feat: assistant evals (#41311)

* chore: bump `supabase` CLI * chore: stricter message types in `generate-v4.ts` * feat: tutorial eval https://www.braintrust.dev/docs/evaluation * feat: project ID for eval * refactor: `generateAssistantResponse` out of `handlePost` * refactor: generateAssistantResponse to lib/ai * feat: factuality eval with assistant response * chore: upgrade braintrust to v1.0.1 * chore: silence tsconfig warning * feat: assertion scorer * fix: aggregate tools across all steps * refactor: strict tool names, remove need for `as const` * refactor: generic tool name type in assertions * feat: transfer mocks from `feature/braintrust` * feat: LLM criteria assertion * feat: braintrust evals workflow * fix: BRAINTRUST_PROJECT_ID * feat: `sql_similar` assertion * fix: `OPENAI_API_KEY` in workflow env * feat: split AssertionScorer into separate scorers * feat: remove tutorial eval * feat: 20 minute CI timeout * feat: category in test case metadata * feat: score with gpt-5 * refactor: dataset to own file, colocate scorers * feat: "gpt-5.2-2025-12-11" for llm as a judge * feat: SQL syntax scorer with `libpg-query` * feat: `evals:setup` and `evals:run` scripts * feat: `evals:setup` in CI * feat: human readable scorer names * chore: rename to "SQL Validity" * feat: add 2 "sql_generation" test cases * feat: update requiredTools in test cases * chore: ignore Cursor MCP config * feat: "Conciseness" score * feat: "Completeness" scorer * fix: generate-v4 test mocks * feat: serialize "steps" for scorer inputs * updated node mem options for typecheck * updated runner * remove ram update as actions handle this * feat: read `BRAINTRUST_PROJECT_ID` from secrets * feat: score helpfulness, remove old scorers * feat: separate `evals:run` and `evals:upload` scripts * feat: passthrough entire classifier result * feat: use live `search_docs` impl, store docs result in metadata * feat: reduce classifier options * feat: filter workflow by `run-evals` PR label or `master` branch * chore: cleanup stubbed mock tools * fix: checkout actual branch with `ref:` * fix: capture search_docs results from all content parts * feat: simplify sql syntax score calculation * feat: use AI SDK's UI message validator * docs: justification for relative `extends` * fix: cleanup leftover validatedMessages * doc: note mock token isn't secret for snyk * fix: mock ui message to pass validation * feat: revert ignoring Cursor MCP config Using `.git/info/exclude` instead until we have an opinion on this * feat: add "tsconfig" as shared-data devDependency, revert relative path in tsconfig * refactor: tool call parsing into function * Update apps/studio/evals/assistant.eval.ts Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> * refactor: organize mock schemas and tool factories --------- Co-authored-by: Ali Waseem <waseema393@gmail.com> Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
2026-05-06 22:18:00 +08:00 · 2025-12-22 23:45:48 -05:00
parent 42a452dad3
commit 072883bcec
15 changed files with 1269 additions and 121 deletions
--- a/.github/workflows/braintrust-evals.yml
+++ b/.github/workflows/braintrust-evals.yml
@@ -0,0 +1,60 @@
+name: Run Braintrust evals
+
+on:
+  push:
+    branches: [master]
+  pull_request:
+    types: [opened, synchronize, labeled]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+permissions:
+  pull-requests: write
+  contents: read
+
+jobs:
+  eval:
+    name: Run evals
+    if: github.event_name == 'push' || (github.event_name == 'pull_request' && contains(github.event.pull_request.labels.*.name, 'run-evals'))
+    runs-on: ubuntu-latest
+    timeout-minutes: 20
+
+    env:
+      BRAINTRUST_PROJECT_ID: ${{ secrets.BRAINTRUST_PROJECT_ID }}
+      OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@08eba0b27e820071cde6df949e0beb9ba4906955 # v4.3.0
+        with:
+          fetch-depth: 0
+          # For PR events, checkout the actual branch so Braintrust can report the correct branch name instead of detached HEAD.
+          # github.head_ref is the PR source branch, github.ref_name is the fallback for push events (e.g., master).
+          ref: ${{ github.head_ref || github.ref_name }}
+
+      - name: Install pnpm
+        uses: pnpm/action-setup@41ff72655975bd51cab0327fa583b6e92b6d3061 # v4.2.0
+        with:
+          run_install: false
+
+      - name: Use Node.js
+        uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        with:
+          node-version-file: ".nvmrc"
+          cache: "pnpm"
+
+      - name: Install Dependencies
+        run: pnpm install --frozen-lockfile
+
+      - name: Setup Evals
+        run: cd apps/studio && pnpm evals:setup
+
+      - name: Run Evals
+        uses: braintrustdata/eval-action@v1
+        with:
+          api_key: ${{ secrets.BRAINTRUST_API_KEY }}
+          runtime: node
+          package_manager: pnpm
+          root: apps/studio
--- a/apps/studio/evals/.gitignore
+++ b/apps/studio/evals/.gitignore
@@ -0,0 +1 @@
+libpg-query.wasm
--- a/apps/studio/evals/assistant.eval.ts
+++ b/apps/studio/evals/assistant.eval.ts
@@ -0,0 +1,123 @@
+import { openai } from '@ai-sdk/openai'
+import { Eval } from 'braintrust'
+import { generateAssistantResponse } from 'lib/ai/generate-assistant-response'
+import { getMockTools } from 'lib/ai/tools/mock-tools'
+import assert from 'node:assert'
+import { dataset } from './dataset'
+import {
+  completenessScorer,
+  concisenessScorer,
+  goalCompletionScorer,
+  sqlSyntaxScorer,
+  toolUsageScorer,
+} from './scorer'
+import { ToolSet, TypedToolCall, TypedToolResult } from 'ai'
+
+assert(process.env.BRAINTRUST_PROJECT_ID, 'BRAINTRUST_PROJECT_ID is not set')
+assert(process.env.OPENAI_API_KEY, 'OPENAI_API_KEY is not set')
+
+Eval('Assistant', {
+  projectId: process.env.BRAINTRUST_PROJECT_ID,
+  data: () => dataset,
+  task: async (input) => {
+    const result = await generateAssistantResponse({
+      model: openai('gpt-5-mini'),
+      messages: [{ id: '1', role: 'user', parts: [{ type: 'text', text: input }] }],
+      tools: await getMockTools(),
+    })
+
+    // `result.toolCalls` only shows the last step, instead aggregate tools across all steps
+    const steps = await result.steps
+
+    const stepsSerialized = steps
+      .map((step) => {
+        const toolCalls = step.toolCalls
+          ?.map((call) => JSON.stringify({ tool: call.toolName, input: call.input }))
+          .join('\n')
+
+        const text = step.text
+        return toolCalls ? `${text}\n${toolCalls}` : text
+      })
+      .join('\n')
+
+    const toolNames: string[] = []
+    const sqlQueries: string[] = []
+    const docs: string[] = []
+
+    for (const step of steps) {
+      for (const [i, toolCall] of step.toolCalls.entries()) {
+        toolNames.push(toolCall.toolName)
+
+        const toolResult = step.toolResults.at(i)
+        if (!toolResult) {
+          continue
+        }
+
+        const parsed = parseToolCall(toolCall, toolResult)
+
+        if (parsed.sqlQuery) {
+          sqlQueries.push(parsed.sqlQuery)
+        }
+        if (parsed.docs) {
+          docs.push(...parsed.docs)
+        }
+      }
+    }
+
+    return {
+      stepsSerialized,
+      toolNames,
+      sqlQueries,
+      docs,
+    }
+  },
+  scores: [
+    toolUsageScorer,
+    sqlSyntaxScorer,
+    goalCompletionScorer,
+    concisenessScorer,
+    completenessScorer,
+  ],
+})
+
+type ParsedToolCall = {
+  /** Query generated by `execute_sql` */
+  sqlQuery?: string
+
+  /** Docs text pulled in from `search_docs` */
+  docs?: string[]
+}
+
+/**
+ * Validate and extract relevant info from a tool call/result
+ */
+function parseToolCall(
+  toolCall: TypedToolCall<ToolSet>,
+  toolResult: TypedToolResult<ToolSet>
+): ParsedToolCall {
+  switch (toolCall.toolName) {
+    case 'execute_sql': {
+      const sqlQuery = toolCall.input.sql
+      if (typeof sqlQuery !== 'string') {
+        return {}
+      }
+
+      return { sqlQuery }
+    }
+    case 'search_docs': {
+      const content = toolResult.output.content
+      if (!content || !Array.isArray(content)) {
+        return {}
+      }
+
+      const docs = content.map((item) => item?.text).filter((text) => typeof text === 'string')
+      if (docs.length === 0) {
+        return {}
+      }
+
+      return { docs }
+    }
+  }
+
+  return {}
+}
--- a/apps/studio/evals/dataset.ts
+++ b/apps/studio/evals/dataset.ts
@@ -0,0 +1,38 @@
+import { AssistantEvalCase } from './scorer'
+
+export const dataset: AssistantEvalCase[] = [
+  {
+    input: 'How do I run WASM in edge functions? Use `search_docs`.',
+    expected: { requiredTools: ['search_docs'] },
+    metadata: { category: ['general_help'] },
+  },
+  {
+    input: 'Check if my project is having issues right now and tell me what to fix first.',
+    expected: {
+      requiredTools: ['get_advisors', 'get_logs'],
+    },
+    metadata: { category: ['debugging', 'rls_policies'] },
+  },
+  {
+    input: 'Create a new table "foods" with columns for "name" and "color"',
+    expected: {
+      requiredTools: ['execute_sql'],
+    },
+    metadata: { category: ['sql_generation', 'schema_design'] },
+  },
+  {
+    input:
+      'Write a SQL query to select all products from the products table where the price is greater than 100',
+    expected: {
+      requiredTools: ['execute_sql'],
+    },
+    metadata: { category: ['sql_generation'] },
+  },
+  {
+    input: 'Create an index on the products table for the name column',
+    expected: {
+      requiredTools: ['execute_sql'],
+    },
+    metadata: { category: ['sql_generation', 'database_optimization'] },
+  },
+]
--- a/apps/studio/evals/scorer.ts
+++ b/apps/studio/evals/scorer.ts
@@ -0,0 +1,159 @@
+import { EvalCase, EvalScorer } from 'braintrust'
+import { LLMClassifierFromTemplate } from 'autoevals'
+import { stripIndent } from 'common-tags'
+import { parse } from 'libpg-query'
+
+const LLM_AS_A_JUDGE_MODEL = 'gpt-5.2-2025-12-11'
+
+type Input = string
+
+type Output = {
+  stepsSerialized: string
+  toolNames: string[]
+  sqlQueries: string[]
+  docs: string[]
+}
+
+export type Expected = {
+  requiredTools?: string[]
+}
+
+// Based on categories in the AssistantMessageRatingSubmittedEvent
+export type AssistantEvalCaseCategory =
+  | 'sql_generation'
+  | 'schema_design'
+  | 'rls_policies'
+  | 'edge_functions'
+  | 'database_optimization'
+  | 'debugging'
+  | 'general_help'
+  | 'other'
+
+export type AssistantEvalCaseMetadata = {
+  category?: AssistantEvalCaseCategory[]
+}
+
+export type AssistantEvalCase = EvalCase<Input, Expected, AssistantEvalCaseMetadata>
+
+export const toolUsageScorer: EvalScorer<Input, Output, Expected> = async ({
+  output,
+  expected,
+}) => {
+  if (!expected.requiredTools) return null
+
+  const presentCount = expected.requiredTools.filter((tool) =>
+    output.toolNames.includes(tool)
+  ).length
+  const totalCount = expected.requiredTools.length
+  const ratio = totalCount === 0 ? 1 : presentCount / totalCount
+
+  return {
+    name: 'Tool Usage',
+    score: ratio,
+  }
+}
+
+export const sqlSyntaxScorer: EvalScorer<Input, Output, Expected> = async ({ output }) => {
+  if (output.sqlQueries === undefined || output.sqlQueries.length === 0) {
+    return null
+  }
+
+  const errors: string[] = []
+  let validQueries = 0
+
+  for (const sql of output.sqlQueries) {
+    try {
+      await parse(sql)
+      validQueries++
+    } catch (error) {
+      const errorMessage = error instanceof Error ? error.message : String(error)
+      errors.push(`SQL syntax error: ${errorMessage}`)
+    }
+  }
+
+  return {
+    name: 'SQL Validity',
+    score: validQueries / output.sqlQueries.length,
+    metadata: errors.length > 0 ? { errors } : undefined,
+  }
+}
+
+const concisenessEvaluator = LLMClassifierFromTemplate<{ input: string }>({
+  name: 'Conciseness',
+  promptTemplate: stripIndent`
+    Evaluate the conciseness of this response.
+
+    Input: {{input}}
+    Output: {{output}}
+
+    Is the response concise and free of unnecessary words?
+    a) Very concise - no wasted words
+    b) Acceptable verbosity - some verbosity but acceptable
+    c) Too verbose - contains superfluous wording or overly verbose
+  `,
+  choiceScores: { a: 1, b: 0.5, c: 0 },
+  useCoT: true,
+  model: LLM_AS_A_JUDGE_MODEL,
+})
+
+export const concisenessScorer: EvalScorer<Input, Output, Expected> = async ({ input, output }) => {
+  return await concisenessEvaluator({
+    input,
+    output: output.stepsSerialized,
+  })
+}
+
+const completenessEvaluator = LLMClassifierFromTemplate<{ input: string }>({
+  name: 'Completeness',
+  promptTemplate: stripIndent`
+    Evaluate whether this response is complete and finished, or if it appears cut off or incomplete.
+
+    Input: {{input}}
+    Output: {{output}}
+
+    Does the response appear complete and finished?
+    a) Complete - response is complete and finished
+    b) Incomplete - response appears cut off, missing parts, or severely incomplete
+  `,
+  choiceScores: { a: 1, b: 0 },
+  useCoT: true,
+  model: LLM_AS_A_JUDGE_MODEL,
+})
+
+export const completenessScorer: EvalScorer<Input, Output, Expected> = async ({
+  input,
+  output,
+}) => {
+  return await completenessEvaluator({
+    input,
+    output: output.stepsSerialized,
+  })
+}
+
+const goalCompletionEvaluator = LLMClassifierFromTemplate<{ input: string }>({
+  name: 'Goal Completion',
+  promptTemplate: stripIndent`
+    Evaluate whether this response addresses what the user asked.
+    
+    Input: {{input}}
+    Output: {{output}}
+    
+    Does the response address what the user asked?
+    a) Fully addresses - completely answers the question or fulfills the request
+    b) Partially addresses - addresses some aspects but misses key parts
+    c) Doesn't address - off-topic or fails to address the request
+  `,
+  choiceScores: { a: 1, b: 0.5, c: 0 },
+  useCoT: true,
+  model: LLM_AS_A_JUDGE_MODEL,
+})
+
+export const goalCompletionScorer: EvalScorer<Input, Output, Expected> = async ({
+  input,
+  output,
+}) => {
+  return await goalCompletionEvaluator({
+    input,
+    output: output.stepsSerialized,
+  })
+}
--- a/apps/studio/lib/ai/generate-assistant-response.ts
+++ b/apps/studio/lib/ai/generate-assistant-response.ts
@@ -0,0 +1,130 @@
+import * as ai from 'ai'
+import {
+  convertToModelMessages,
+  isToolUIPart,
+  type LanguageModel,
+  type ModelMessage,
+  stepCountIs,
+  type ToolSet,
+  type UIMessage,
+} from 'ai'
+import { wrapAISDK } from 'braintrust'
+import { source } from 'common-tags'
+
+import type { AiOptInLevel } from 'hooks/misc/useOrgOptedIntoAi'
+import {
+  CHAT_PROMPT,
+  EDGE_FUNCTION_PROMPT,
+  GENERAL_PROMPT,
+  PG_BEST_PRACTICES,
+  RLS_PROMPT,
+  REALTIME_PROMPT,
+  SECURITY_PROMPT,
+  LIMITATIONS_PROMPT,
+} from 'lib/ai/prompts'
+import { sanitizeMessagePart } from 'lib/ai/tools/tool-sanitizer'
+
+const { streamText } = wrapAISDK(ai)
+
+export async function generateAssistantResponse({
+  messages: rawMessages,
+  model,
+  tools,
+  aiOptInLevel = 'schema',
+  getSchemas,
+  projectRef,
+  chatName,
+  promptProviderOptions,
+  providerOptions,
+  abortSignal,
+}: {
+  messages: UIMessage[]
+  model: LanguageModel
+  tools: ToolSet
+  aiOptInLevel?: AiOptInLevel
+  getSchemas?: () => Promise<string>
+  projectRef?: string
+  chatName?: string
+  promptProviderOptions?: Record<string, any>
+  providerOptions?: Record<string, any>
+  abortSignal?: AbortSignal
+}) {
+  // Only returns last 7 messages
+  // Filters out tools with invalid states
+  // Filters out tool outputs based on opt-in level using renderingToolOutputParser
+  const messages = (rawMessages || []).slice(-7).map((msg) => {
+    if (msg && msg.role === 'assistant' && 'results' in msg) {
+      const cleanedMsg = { ...msg }
+      delete cleanedMsg.results
+      return cleanedMsg
+    }
+    if (msg && msg.role === 'assistant' && msg.parts) {
+      const cleanedParts = msg.parts
+        .filter((part) => {
+          if (isToolUIPart(part)) {
+            const invalidStates = ['input-streaming', 'input-available', 'output-error']
+            return !invalidStates.includes(part.state)
+          }
+          return true
+        })
+        .map((part) => {
+          return sanitizeMessagePart(part, aiOptInLevel)
+        })
+      return { ...msg, parts: cleanedParts }
+    }
+    return msg
+  })
+
+  const schemasString =
+    aiOptInLevel !== 'disabled' && getSchemas
+      ? await getSchemas()
+      : "You don't have access to any schemas."
+
+  // Important: do not use dynamic content in the system prompt or Bedrock will not cache it
+  const system = source`
+    ${GENERAL_PROMPT}
+    ${CHAT_PROMPT}
+    ${PG_BEST_PRACTICES}
+    ${RLS_PROMPT}
+    ${EDGE_FUNCTION_PROMPT}
+    ${REALTIME_PROMPT}
+    ${SECURITY_PROMPT}
+    ${LIMITATIONS_PROMPT}
+  `
+
+  // Note: these must be of type `CoreMessage` to prevent AI SDK from stripping `providerOptions`
+  // https://github.com/vercel/ai/blob/81ef2511311e8af34d75e37fc8204a82e775e8c3/packages/ai/core/prompt/standardize-prompt.ts#L83-L88
+  const assistantContent =
+    projectRef || chatName || schemasString !== "You don't have access to any schemas."
+      ? `The user's current project is ${projectRef || 'unknown'}. Their available schemas are: ${schemasString}. The current chat name is: ${chatName || 'unnamed'}`
+      : undefined
+
+  const coreMessages: ModelMessage[] = [
+    {
+      role: 'system',
+      content: system,
+      ...(promptProviderOptions && {
+        providerOptions: promptProviderOptions,
+      }),
+    },
+    ...(assistantContent
+      ? [
+          {
+            role: 'assistant' as const,
+            // Add any dynamic context here
+            content: assistantContent,
+          },
+        ]
+      : []),
+    ...convertToModelMessages(messages),
+  ]
+
+  return streamText({
+    model,
+    stopWhen: stepCountIs(5),
+    messages: coreMessages,
+    ...(providerOptions && { providerOptions }),
+    tools,
+    ...(abortSignal && { abortSignal }),
+  })
+}
--- a/apps/studio/lib/ai/tools/mock-tools.ts
+++ b/apps/studio/lib/ai/tools/mock-tools.ts
@@ -0,0 +1,324 @@
+import { tool, type ToolSet } from 'ai'
+import { getRenderingTools } from '../tools/rendering-tools'
+import { z } from 'zod'
+import { getMcpTools } from 'lib/ai/tools/mcp-tools'
+import assert from 'node:assert'
+
+const listTablesInputSchema = z.object({
+  schemas: z.array(z.string()).describe('The schema names to list.'),
+})
+
+const getAdvisorsInputSchema = z.object({
+  type: z.enum(['security', 'performance']).optional(),
+})
+
+const getLogsInputSchema = z.object({
+  limit: z.number().min(1).max(100).optional(),
+  level: z.enum(['debug', 'info', 'warning', 'error']).optional(),
+  source: z.enum(['postgres', 'auth', 'storage', 'edge_function']).optional(),
+  search: z.string().optional(),
+})
+
+const listPoliciesInputSchema = z.object({
+  schemas: z.array(z.string()).describe('The schema names to get the policies for'),
+})
+
+const MOCK_TABLES_DATA = [
+  {
+    name: 'user_documents',
+    rls_enabled: false,
+    columns: [
+      { name: 'id', data_type: 'bigint' },
+      { name: 'user_id', data_type: 'uuid' },
+      { name: 'title', data_type: 'text' },
+    ],
+  },
+  {
+    name: 'customers',
+    rls_enabled: true,
+    columns: [
+      { name: 'id', data_type: 'uuid' },
+      { name: 'tenant_id', data_type: 'uuid' },
+      { name: 'email', data_type: 'text' },
+    ],
+  },
+  {
+    name: 'projects',
+    rls_enabled: false,
+    columns: [
+      { name: 'id', data_type: 'uuid' },
+      { name: 'organization_id', data_type: 'uuid' },
+      { name: 'name', data_type: 'text' },
+    ],
+  },
+  {
+    name: 'user_organizations',
+    rls_enabled: true,
+    columns: [
+      { name: 'user_id', data_type: 'uuid' },
+      { name: 'organization_id', data_type: 'uuid' },
+    ],
+  },
+]
+
+const MOCK_EXTENSIONS_DATA = [
+  { name: 'pgcrypto', schema: 'extensions', installed_version: '1.3' },
+  { name: 'uuid-ossp', schema: 'extensions', installed_version: '1.1' },
+]
+
+const MOCK_EDGE_FUNCTIONS_DATA = [
+  { name: 'hello-world', last_deployed_at: '2024-06-10T12:30:00Z' },
+  { name: 'daily-metrics-sync', last_deployed_at: '2024-06-18T08:15:00Z' },
+  { name: 'select-from-table-with-auth-rls', last_deployed_at: '2024-06-19T09:20:00Z' },
+]
+
+const MOCK_ADVISORIES_DATA = [
+  {
+    id: '0016_materialized_view_in_api',
+    level: 'warning',
+    category: 'security',
+    message: 'Materialized views in API schema can bypass RLS. Move them to private schema.',
+    remediationUrl:
+      'https://supabase.com/docs/guides/database/database-advisors?queryGroups=lint&lint=0016_materialized_view_in_api',
+  },
+  {
+    id: '0031_functions_no_rls_guard',
+    level: 'notice',
+    category: 'security',
+    message: 'Function api.health_check should verify auth context before querying tables.',
+    remediationUrl:
+      'https://supabase.com/docs/guides/database/database-advisors?queryGroups=lint&lint=0031_functions_no_rls_guard',
+  },
+  {
+    id: '1012_slow_query',
+    level: 'info',
+    category: 'performance',
+    message:
+      'Query on table edge_function_logs exceeded 3s average execution time over the last hour.',
+    remediationUrl: 'https://supabase.com/docs/guides/platform/performance-advisors#slow-queries',
+  },
+]
+
+const MOCK_LOGS_DATA = [
+  {
+    id: 'log-001',
+    timestamp: '2024-06-20T14:12:00Z',
+    level: 'error',
+    source: 'edge_function' as const,
+    target: 'hello-world',
+    message: "TypeError: fetch failed at await supabase.functions.invoke('analytics')",
+  },
+  {
+    id: 'log-002',
+    timestamp: '2024-06-20T14:05:30Z',
+    level: 'warning',
+    source: 'postgres' as const,
+    target: 'connection_pool',
+    message: 'Query timeout exceeded for statement SELECT * FROM public.audit_log_entries',
+  },
+  {
+    id: 'log-003',
+    timestamp: '2024-06-20T13:59:10Z',
+    level: 'info',
+    source: 'edge_function' as const,
+    target: 'daily-metrics-sync',
+    message: 'Invocation completed in 520ms',
+  },
+  {
+    id: 'log-004',
+    timestamp: '2024-06-20T13:50:00Z',
+    level: 'error',
+    source: 'postgres' as const,
+    target: 'trigger:refresh_materialized_views',
+    message: 'permission denied for relation user_documents',
+  },
+  {
+    id: 'log-005',
+    timestamp: '2024-06-20T13:45:00Z',
+    level: 'info',
+    source: 'auth' as const,
+    target: 'email-confirmation',
+    message: 'Sent verification email to alex@example.com',
+  },
+]
+
+function createMockedRenderingTools() {
+  const renderingTools = getRenderingTools()
+
+  return Object.fromEntries(
+    Object.entries(renderingTools).map(([name, baseTool]) => {
+      if (typeof baseTool.execute === 'function') {
+        return [name, baseTool]
+      }
+
+      const statusMessage =
+        name === 'execute_sql'
+          ? 'SQL execution mocked successfully.'
+          : name === 'deploy_edge_function'
+            ? 'Edge Function deployment mocked successfully.'
+            : 'Tool call mocked successfully.'
+
+      return [
+        name,
+        {
+          ...baseTool,
+          execute: async () => ({ status: statusMessage }),
+        },
+      ]
+    })
+  ) as typeof renderingTools
+}
+
+function createMockListTablesTool() {
+  return tool({
+    description: 'Lists tables and columns for the provided schemas.',
+    inputSchema: listTablesInputSchema,
+    execute: async ({ schemas }: { schemas: string[] }) => {
+      const effectiveSchemas = schemas?.length ? schemas : ['public']
+      return effectiveSchemas.map((schema) => ({
+        schema,
+        tables: MOCK_TABLES_DATA,
+      }))
+    },
+  })
+}
+
+function createMockListExtensionsTool() {
+  return tool({
+    description: 'Lists installed database extensions.',
+    inputSchema: z.object({}),
+    execute: async () => {
+      return MOCK_EXTENSIONS_DATA
+    },
+  })
+}
+
+function createMockListEdgeFunctionsTool() {
+  return tool({
+    description: 'Lists available Supabase Edge Functions.',
+    inputSchema: z.object({}),
+    execute: async () => {
+      return MOCK_EDGE_FUNCTIONS_DATA
+    },
+  })
+}
+
+function createMockGetAdvisorsTool() {
+  return tool({
+    description: 'Returns advisory notices for the project (mocked).',
+    inputSchema: getAdvisorsInputSchema,
+    execute: async ({ type }: { type?: 'security' | 'performance' }) => {
+      if (type) {
+        return MOCK_ADVISORIES_DATA.filter((advisory) => advisory.category === type)
+      }
+      return MOCK_ADVISORIES_DATA
+    },
+  })
+}
+
+function createMockGetLogsTool() {
+  return tool({
+    description: 'Fetches recent project logs for debugging or health checks (mocked).',
+    inputSchema: getLogsInputSchema,
+    execute: async ({
+      limit = 10,
+      level,
+      source,
+      search,
+    }: {
+      limit?: number
+      level?: 'debug' | 'info' | 'warning' | 'error'
+      source?: 'postgres' | 'auth' | 'storage' | 'edge_function'
+      search?: string
+    }) => {
+      let filtered = MOCK_LOGS_DATA
+
+      if (level) {
+        filtered = filtered.filter((entry) => entry.level === level)
+      }
+
+      if (source) {
+        filtered = filtered.filter((entry) => entry.source === source)
+      }
+
+      if (search) {
+        const needle = search.toLowerCase()
+        filtered = filtered.filter((entry) =>
+          `${entry.message} ${entry.target}`.toLowerCase().includes(needle)
+        )
+      }
+
+      return filtered.slice(0, limit)
+    },
+  })
+}
+
+function createMockListPoliciesTool() {
+  return tool({
+    description: 'Get existing RLS policies for provided schemas.',
+    inputSchema: listPoliciesInputSchema,
+    execute: async ({ schemas }: { schemas: string[] }) => {
+      const effectiveSchemas = schemas?.length ? schemas : ['public']
+      const results = [] as Array<{
+        schema: string
+        table: string
+        policies: Array<{
+          name: string
+          command: 'select' | 'insert' | 'update' | 'delete'
+          using?: string
+          check?: string
+        }>
+      }>
+
+      for (const schema of effectiveSchemas) {
+        if (schema !== 'public') continue
+        results.push(
+          {
+            schema,
+            table: 'customers',
+            policies: [
+              {
+                name: 'customers_tenant_select',
+                command: 'select',
+                using: "(auth.jwt() ->> 'tenant_id')::uuid = tenant_id",
+              },
+            ],
+          },
+          { schema, table: 'user_documents', policies: [] },
+          { schema, table: 'projects', policies: [] }
+        )
+      }
+      return results
+    },
+  })
+}
+
+/**
+ * Deterministic mock implementations of MCP/platform tools for evals.
+ * These mirror tool names used in prompts so the model can call them,
+ * but return stable, static data for repeatable tests.
+ *
+ * Note: search_docs uses the real implementation
+ */
+export async function getMockTools() {
+  const mockedRenderingTools = createMockedRenderingTools()
+
+  const { search_docs } = await getMcpTools({
+    accessToken: 'mock-access-token',
+    projectRef: 'mock-project-ref',
+    aiOptInLevel: 'schema_and_log_and_data',
+  })
+
+  assert(search_docs, 'search_docs tool not available from MCP server')
+
+  return {
+    ...mockedRenderingTools,
+    search_docs,
+    list_tables: createMockListTablesTool(),
+    list_extensions: createMockListExtensionsTool(),
+    list_edge_functions: createMockListEdgeFunctionsTool(),
+    get_advisors: createMockGetAdvisorsTool(),
+    get_logs: createMockGetLogsTool(),
+    list_policies: createMockListPoliciesTool(),
+  }
+}
--- a/apps/studio/lib/api/generate-v4.test.ts
+++ b/apps/studio/lib/api/generate-v4.test.ts
@@ -2,9 +2,10 @@ import { expect, test, vi } from 'vitest'
 // End of third-party imports

 import generateV4 from '../../pages/api/ai/sql/generate-v4'
-import { sanitizeMessagePart } from '../ai/tools/tool-sanitizer'
+import { sanitizeMessagePart } from 'lib/ai/tools/tool-sanitizer'
+import { UIMessage } from 'ai'

-vi.mock('../ai/tools/tool-sanitizer', () => ({
+vi.mock('lib/ai/tools/tool-sanitizer', () => ({
  sanitizeMessagePart: vi.fn((part) => part),
 }))

@@ -17,20 +18,24 @@ test('generateV4 calls the tool sanitizer', async () => {
    body: {
      messages: [
        {
+          id: 'test-msg-id',
          role: 'assistant',
          parts: [
            {
              type: 'tool-execute_sql',
              state: 'output-available',
-              output: 'test output',
+              toolCallId: 'test-tool-call-id',
+              input: { sql: 'SELECT * FROM users' },
+              output: [{ id: 1, name: 'test-output' }],
            },
          ],
        },
-      ],
+      ] satisfies UIMessage[],
      projectRef: 'test-project',
      connectionString: 'test-connection',
      orgSlug: 'test-org',
    },
+    on: vi.fn(),
  }

  const mockRes = {
@@ -63,13 +68,15 @@ test('generateV4 calls the tool sanitizer', async () => {
    getTools: vi.fn().mockResolvedValue({}),
  }))

-  vi.mock('ai', () => ({
-    streamText: vi.fn().mockReturnValue({
-      pipeUIMessageStreamToResponse: vi.fn(),
-    }),
-    convertToModelMessages: vi.fn((msgs) => msgs),
-    stepCountIs: vi.fn(),
-  }))
+  vi.mock('ai', async () => {
+    const actual = await vi.importActual('ai')
+    return {
+      ...actual,
+      streamText: vi.fn().mockReturnValue({
+        pipeUIMessageStreamToResponse: vi.fn(),
+      }),
+    }
+  })

  await generateV4(mockReq as any, mockRes as any)

--- a/apps/studio/package.json
+++ b/apps/studio/package.json
@@ -23,7 +23,10 @@
    "prettier:write": "prettier --cache --write .",
    "build:deno-types": "tsx scripts/deno-types.ts",
    "build:graphql-types": "tsx scripts/download-graphql-schema.mts && pnpm graphql-codegen --config scripts/codegen.ts",
-    "build:graphql-types:watch": "pnpm graphql-codegen --config scripts/codegen.ts --watch"
+    "build:graphql-types:watch": "pnpm graphql-codegen --config scripts/codegen.ts --watch",
+    "evals:setup": "cp node_modules/libpg-query/wasm/libpg-query.wasm evals/libpg-query.wasm",
+    "evals:run": "braintrust eval --no-send-logs evals/assistant.eval.ts",
+    "evals:upload": "braintrust eval evals/assistant.eval.ts"
  },
  "dependencies": {
    "@ai-sdk/amazon-bedrock": "^3.0.0",
@@ -183,6 +186,8 @@
    "@vitest/coverage-v8": "^3.2.0",
    "@vitest/ui": "^3.2.0",
    "api-types": "workspace:*",
+    "autoevals": "^0.0.131",
+    "braintrust": "^1.0.2",
    "common": "workspace:*",
    "config": "workspace:*",
    "date-fns": "^2.30.0",
@@ -191,6 +196,7 @@
    "graphql-ws": "5.14.1",
    "import-in-the-middle": "^2.0.0",
    "jsdom-testing-mocks": "^1.13.1",
+    "libpg-query": "17.6.0",
    "msw": "^2.3.0",
    "next-router-mock": "^0.9.13",
    "node-mocks-http": "^1.17.2",
--- a/apps/studio/pages/api/ai/sql/generate-v4.ts
+++ b/apps/studio/pages/api/ai/sql/generate-v4.ts
@@ -1,6 +1,5 @@
 import pgMeta from '@supabase/pg-meta'
-import { convertToModelMessages, type ModelMessage, stepCountIs, streamText } from 'ai'
-import { source } from 'common-tags'
+import { safeValidateUIMessages } from 'ai'
 import type { NextApiRequest, NextApiResponse } from 'next'
 import z from 'zod'

@@ -9,18 +8,8 @@ import { executeSql } from 'data/sql/execute-sql-query'
 import type { AiOptInLevel } from 'hooks/misc/useOrgOptedIntoAi'
 import { getModel } from 'lib/ai/model'
 import { getOrgAIDetails } from 'lib/ai/org-ai-details'
-import {
-  CHAT_PROMPT,
-  EDGE_FUNCTION_PROMPT,
-  GENERAL_PROMPT,
-  PG_BEST_PRACTICES,
-  RLS_PROMPT,
-  REALTIME_PROMPT,
-  SECURITY_PROMPT,
-  LIMITATIONS_PROMPT,
-} from 'lib/ai/prompts'
+import { generateAssistantResponse } from 'lib/ai/generate-assistant-response'
 import { getTools } from 'lib/ai/tools'
-import { sanitizeMessagePart } from 'lib/ai/tools/tool-sanitizer'
 import apiWrapper from 'lib/api/apiWrapper'
 import { executeQuery } from 'lib/api/self-hosted/query'

@@ -89,6 +78,14 @@ async function handlePost(req: NextApiRequest, res: NextApiResponse) {
    model: requestedModel,
  } = data

+  const messagesValidation = await safeValidateUIMessages({ messages: rawMessages })
+  if (!messagesValidation.success) {
+    return res
+      .status(400)
+      .json({ error: 'Invalid request body', message: messagesValidation.error.message })
+  }
+  const messages = messagesValidation.data
+
  let aiOptInLevel: AiOptInLevel = 'disabled'
  let isLimited = false

@@ -114,32 +111,6 @@ async function handlePost(req: NextApiRequest, res: NextApiResponse) {
    }
  }

-  // Only returns last 7 messages
-  // Filters out tools with invalid states
-  // Filters out tool outputs based on opt-in level using renderingToolOutputParser
-  const messages = (rawMessages || []).slice(-7).map((msg: any) => {
-    if (msg && msg.role === 'assistant' && 'results' in msg) {
-      const cleanedMsg = { ...msg }
-      delete cleanedMsg.results
-      return cleanedMsg
-    }
-    if (msg && msg.role === 'assistant' && msg.parts) {
-      const cleanedParts = msg.parts
-        .filter((part: any) => {
-          if (part.type.startsWith('tool-')) {
-            const invalidStates = ['input-streaming', 'input-available', 'output-error']
-            return !invalidStates.includes(part.state)
-          }
-          return true
-        })
-        .map((part: any) => {
-          return sanitizeMessagePart(part, aiOptInLevel)
-        })
-      return { ...msg, parts: cleanedParts }
-    }
-    return msg
-  })
-
  const {
    model,
    error: modelError,
@@ -157,67 +128,10 @@ async function handlePost(req: NextApiRequest, res: NextApiResponse) {
  }

  try {
-    // Get a list of all schemas to add to context
-    const pgMetaSchemasList = pgMeta.schemas.list()
-    type Schemas = z.infer<(typeof pgMetaSchemasList)['zod']>
-
-    const { result: schemas } =
-      aiOptInLevel !== 'disabled'
-        ? await executeSql<Schemas>(
-            {
-              projectRef,
-              connectionString,
-              sql: pgMetaSchemasList.sql,
-            },
-            undefined,
-            {
-              'Content-Type': 'application/json',
-              ...(authorization && { Authorization: authorization }),
-            },
-            IS_PLATFORM ? undefined : executeQuery
-          )
-        : { result: [] }
-
-    const schemasString =
-      schemas?.length > 0
-        ? `The available database schema names are: ${JSON.stringify(schemas)}`
-        : "You don't have access to any schemas."
-
-    // Important: do not use dynamic content in the system prompt or Bedrock will not cache it
-    const system = source`
-      ${GENERAL_PROMPT}
-      ${CHAT_PROMPT}
-      ${PG_BEST_PRACTICES}
-      ${RLS_PROMPT}
-      ${EDGE_FUNCTION_PROMPT}
-      ${REALTIME_PROMPT}
-      ${SECURITY_PROMPT}
-      ${LIMITATIONS_PROMPT}
-    `
-
-    // Note: these must be of type `CoreMessage` to prevent AI SDK from stripping `providerOptions`
-    // https://github.com/vercel/ai/blob/81ef2511311e8af34d75e37fc8204a82e775e8c3/packages/ai/core/prompt/standardize-prompt.ts#L83-L88
-    const coreMessages: ModelMessage[] = [
-      {
-        role: 'system',
-        content: system,
-        ...(promptProviderOptions && {
-          providerOptions: promptProviderOptions,
-        }),
-      },
-      {
-        role: 'assistant',
-        // Add any dynamic context here
-        content: `The user's current project is ${projectRef}. Their available schemas are: ${schemasString}. The current chat name is: ${chatName}`,
-      },
-      ...convertToModelMessages(messages),
-    ]
-
    const abortController = new AbortController()
    req.on('close', () => abortController.abort())
    req.on('aborted', () => abortController.abort())

-    // Get tools
    const tools = await getTools({
      projectRef,
      connectionString,
@@ -226,12 +140,40 @@ async function handlePost(req: NextApiRequest, res: NextApiResponse) {
      accessToken,
    })

-    const result = streamText({
+    // Get a list of all schemas to add to context
+    const getSchemas = async (): Promise<string> => {
+      const pgMetaSchemasList = pgMeta.schemas.list()
+      type Schemas = z.infer<(typeof pgMetaSchemasList)['zod']>
+
+      const { result: schemas } = await executeSql<Schemas>(
+        {
+          projectRef,
+          connectionString,
+          sql: pgMetaSchemasList.sql,
+        },
+        undefined,
+        {
+          'Content-Type': 'application/json',
+          ...(authorization && { Authorization: authorization }),
+        },
+        IS_PLATFORM ? undefined : executeQuery
+      )
+
+      return schemas?.length > 0
+        ? `The available database schema names are: ${JSON.stringify(schemas)}`
+        : "You don't have access to any schemas."
+    }
+
+    const result = await generateAssistantResponse({
+      messages,
      model,
-      stopWhen: stepCountIs(5),
-      messages: coreMessages,
-      ...(providerOptions && { providerOptions }),
      tools,
+      aiOptInLevel,
+      getSchemas: aiOptInLevel !== 'disabled' ? getSchemas : undefined,
+      projectRef,
+      chatName,
+      promptProviderOptions,
+      providerOptions,
      abortSignal: abortController.signal,
    })

--- a/package.json
+++ b/package.json
@@ -49,7 +49,7 @@
    "prettier-plugin-sql-cst": "^0.11.0",
    "rimraf": "^6.0.0",
    "sass": "^1.72.0",
-    "supabase": "^2.58.5",
+    "supabase": "^2.65.6",
    "supports-color": "^8.0.0",
    "tailwindcss": "catalog:",
    "turbo": "2.3.3",
--- a/packages/shared-data/package.json
+++ b/packages/shared-data/package.json
@@ -8,6 +8,9 @@
    "preinstall": "npx only-allow pnpm",
    "clean": "rimraf node_modules"
  },
+  "devDependencies": {
+    "tsconfig": "workspace:"
+  },
  "author": "",
  "license": "MIT"
 }
--- a/pnpm-lock.yaml
+++ b/pnpm-lock.yaml
@@ -6,12 +6,63 @@ settings:

 catalogs:
  default:
+    '@sentry/nextjs':
+      specifier: ^10.26.0
+      version: 10.27.0
+    '@supabase/auth-js':
+      specifier: 2.87.0
+      version: 2.87.0
+    '@supabase/postgrest-js':
+      specifier: 2.87.0
+      version: 2.87.0
+    '@supabase/realtime-js':
+      specifier: 2.87.0
+      version: 2.87.0
    '@supabase/supabase-js':
      specifier: 2.87.0
      version: 2.87.0
+    '@types/node':
+      specifier: ^22.0.0
+      version: 22.13.14
+    '@types/react':
+      specifier: ^18.3.0
+      version: 18.3.3
+    '@types/react-dom':
+      specifier: ^18.3.0
+      version: 18.3.0
+    next:
+      specifier: ^15.5.9
+      version: 15.5.9
+    react:
+      specifier: ^18.3.0
+      version: 18.3.1
+    react-dom:
+      specifier: ^18.3.0
+      version: 18.3.1
+    recharts:
+      specifier: ^2.15.4
+      version: 2.15.4
+    tailwindcss:
+      specifier: 3.4.1
+      version: 3.4.1
    tsx:
      specifier: 4.20.3
      version: 4.20.3
+    typescript:
+      specifier: ~5.9.0
+      version: 5.9.2
+    valtio:
+      specifier: ^1.12.0
+      version: 1.12.0
+    vite:
+      specifier: ^7.1.11
+      version: 7.1.11
+    vitest:
+      specifier: ^3.2.0
+      version: 3.2.4
+    zod:
+      specifier: ^3.25.76
+      version: 3.25.76

 overrides:
  '@eslint/eslintrc>js-yaml': ^4.1.1
@@ -52,8 +103,8 @@ importers:
        specifier: ^1.72.0
        version: 1.72.0
      supabase:
-        specifier: ^2.58.5
-        version: 2.58.5(supports-color@8.1.1)
+        specifier: ^2.65.6
+        version: 2.67.1(supports-color@8.1.1)
      supports-color:
        specifier: ^8.0.0
        version: 8.1.1
@@ -489,7 +540,7 @@ importers:
        version: 0.3.0(react-dom@18.3.1(react@18.3.1))(react@18.3.1)
      nuqs:
        specifier: ^1.19.1
-        version: 1.19.1(next@15.5.9(@opentelemetry/api@1.9.0)(@playwright/test@1.56.1)(react-dom@18.3.1(react@18.3.1))(react@18.3.1)(sass@1.77.4))
+        version: 1.19.1(next@15.5.9(@babel/core@7.28.4(supports-color@8.1.1))(@opentelemetry/api@1.9.0)(@playwright/test@1.56.1)(react-dom@18.3.1(react@18.3.1))(react@18.3.1)(sass@1.77.4))
      openai:
        specifier: ^4.75.1
        version: 4.104.0(encoding@0.1.13)(ws@8.18.3)(zod@3.25.76)
@@ -1170,6 +1221,12 @@ importers:
      api-types:
        specifier: workspace:*
        version: link:../../packages/api-types
+      autoevals:
+        specifier: ^0.0.131
+        version: 0.0.131(encoding@0.1.13)(ws@8.18.3)
+      braintrust:
+        specifier: ^1.0.2
+        version: 1.0.2(@aws-sdk/credential-provider-web-identity@3.830.0)(supports-color@8.1.1)(zod@3.25.76)
      date-fns:
        specifier: ^2.30.0
        version: 2.30.0
@@ -1188,6 +1245,9 @@ importers:
      jsdom-testing-mocks:
        specifier: ^1.13.1
        version: 1.13.1
+      libpg-query:
+        specifier: 17.6.0
+        version: 17.6.0
      msw:
        specifier: ^2.3.0
        version: 2.11.3(@types/node@22.13.14)(typescript@5.9.2)
@@ -2173,7 +2233,11 @@ importers:
        specifier: 'catalog:'
        version: 3.2.4(@types/node@22.13.14)(@vitest/ui@3.2.4)(jiti@2.5.1)(jsdom@20.0.3(supports-color@8.1.1))(msw@2.11.3(@types/node@22.13.14)(typescript@5.9.2))(sass@1.77.4)(supports-color@8.1.1)(terser@5.39.0)(tsx@4.20.3)(yaml@2.8.1)

-  packages/shared-data: {}
+  packages/shared-data:
+    devDependencies:
+      tsconfig:
+        specifier: 'workspace:'
+        version: link:../tsconfig

  packages/tsconfig: {}

@@ -2671,6 +2735,10 @@ packages:
    peerDependencies:
      zod: ^3.25.76 || ^4

+  '@ai-sdk/provider@1.1.3':
+    resolution: {integrity: sha512-qZMxYJ0qqX/RfnuIaab+zp8UAeJn/ygXXAffR5I4N0n1IrvA6qBsjc8hXLmBiMV2zoXlifkacF7sEFnYnjBcqg==}
+    engines: {node: '>=18'}
+
  '@ai-sdk/provider@2.0.0':
    resolution: {integrity: sha512-6o7Y2SeO9vFKB8lArHXehNuusnpddKPk7xqL7T2/b+OvXMRIXUO1rR4wcv1hAFUAT9avGZshty3Wlua/XA7TvA==}
    engines: {node: '>=18'}
@@ -3244,6 +3312,10 @@ packages:
      react: ^16.8.0 || ^17 || ^18 || ^19
      react-dom: ^16.8.0 || ^17 || ^18 || ^19

+  '@colors/colors@1.5.0':
+    resolution: {integrity: sha512-ooWCrlZP11i8GImSjTHYHLkvFDP48nS4+204nGb1RiX/WXYHmJA2III9/e2DWVabCESdW7hBAEzHRqUn9OUVvQ==}
+    engines: {node: '>=0.1.90'}
+
  '@contentlayer2/cli@0.4.3':
    resolution: {integrity: sha512-ZJ+Iiu2rVI50x60XoqnrsO/Q8eqFX5AlP1L0U/3ygaAas3tnOqTzQZ1UsxYQMpJzcLok24ddlhKfQKbCMUJPiQ==}

@@ -4916,6 +4988,9 @@ packages:
  '@next/bundle-analyzer@16.0.4':
    resolution: {integrity: sha512-6IajJ23QrXW5RTJj2lRHcBM8mxcEl+vgd7XXVODQG/BcJyjgIP1k5OefdRl+P80btPvHeHoV4fIgC1so25pXcg==}

+  '@next/env@14.2.35':
+    resolution: {integrity: sha512-DuhvCtj4t9Gwrx80dmz2F4t/zKQ4ktN8WrMwOuVzkJfBilwAwGr6v16M5eI8yCuZ63H9TTuEU09Iu2HqkzFPVQ==}
+
  '@next/env@15.5.9':
    resolution: {integrity: sha512-4GlTZ+EJM7WaW2HEZcyU317tIQDjkQIyENDLxYJfSWlfqguN+dHkZgyQTV/7ykvobU7yEH5gKvreNrH4B6QgIg==}

@@ -6190,6 +6265,9 @@ packages:
  '@pgsql/types@15.0.2':
    resolution: {integrity: sha512-K3gtnbqbSUuUVmPm143qx5Gy2EmKuooshV95yMD48EUQ1256sgZBriEfY61OWJnlzdREdqHTIOxQqpZAb7XdZg==}

+  '@pgsql/types@17.6.2':
+    resolution: {integrity: sha512-1UtbELdbqNdyOShhrVfSz3a1gDi0s9XXiQemx+6QqtsrXe62a6zOGU+vjb2GRfG5jeEokI1zBBcfD42enRv0Rw==}
+
  '@pkgjs/parseargs@0.11.0':
    resolution: {integrity: sha512-+1VkjdD0QBLPodGrJUeqarH8VAIvQODIbwh9XpP5Syisf7YoQgsJKPNFoqqLQlu+VQ/tVSshMR6loPMn8U+dPg==}
    engines: {node: '>=14'}
@@ -9749,6 +9827,15 @@ packages:
  '@usercentrics/cmp-browser-sdk@4.42.0':
    resolution: {integrity: sha512-/cik01TdeiYUV1+EasK83Ip05TDqmowM5tmWGfRDdaOrneSic5BedBdJHKya4pP2vwiYAtzDYVJe4BwPt2M16g==}

+  '@vercel/functions@1.6.0':
+    resolution: {integrity: sha512-R6FKQrYT5MZs5IE1SqeCJWxMuBdHawFcCZboKKw8p7s+6/mcd55Gx6tWmyKnQTyrSEA04NH73Tc9CbqpEle8RA==}
+    engines: {node: '>= 16'}
+    peerDependencies:
+      '@aws-sdk/credential-provider-web-identity': '*'
+    peerDependenciesMeta:
+      '@aws-sdk/credential-provider-web-identity':
+        optional: true
+
  '@vercel/functions@2.1.0':
    resolution: {integrity: sha512-1gSbK9zfrbJxk1JTBVERDhLi01mK3fz+gw4GjOjZwHnqs0zsBhQA70HGVtXQX/Z3BTRMfbpAEMVDfhecRw0lDA==}
    engines: {node: '>= 18'}
@@ -10358,6 +10445,9 @@ packages:
    resolution: {integrity: sha512-Hdw8qdNiqdJ8LqT0iK0sVzkFbzg6fhnQqqfWhBDxcHZvU75+B+ayzTy8x+k5Ix0Y92XOhOUlx74ps+bA6BeYMQ==}
    engines: {node: '>=8'}

+  autoevals@0.0.131:
+    resolution: {integrity: sha512-F+3lraja+Ms7n1M2cpWl65N7AYx4sPocRW454H5HlSGabYMfuFOUxw8IXmEYDkQ38BxtZ0Wd5ZAQj9RF59YJWw==}
+
  autolinker@0.28.1:
    resolution: {integrity: sha512-zQAFO1Dlsn69eXaO6+7YZc+v84aquQKbwpzCE3L0stj56ERn9hutFxPopViLjo9G+rWwjozRhgS5KJ25Xy19cQ==}

@@ -10458,6 +10548,9 @@ packages:
    resolution: {integrity: sha512-jDctJ/IVQbZoJykoeHbhXpOlNBqGNcwXJKJog42E5HDPUwQTSdjCHdihjj0DlnheQ7blbT6dHOafNAiS8ooQKA==}
    engines: {node: '>=8'}

+  binary-search@1.3.6:
+    resolution: {integrity: sha512-nbE1WxOTTrUWIfsfZ4aHGYu5DOuNkbxGokjV6Z2kxfJK3uaAb8zNK1muzOeipoLHZjInT4Br88BHpzevc681xA==}
+
  bindings@1.5.0:
    resolution: {integrity: sha512-p2q/t/mhvuOj/UeLlV6566GD/guowlr0hHxClI0W9m7MWYkL1F0hLo+0Aexs9HSPCtR1SXQ0TD3MMKrXZajbiQ==}

@@ -10491,6 +10584,10 @@ packages:
    resolution: {integrity: sha512-2hCgjEmP8YLWQ130n2FerGv7rYpfBmnmp9Uy2Le1vge6X3gZIfSmEzP5QTDElFxcvVcXlEn8Aq6MU/PZygIOog==}
    engines: {node: '>=14.16'}

+  boxen@8.0.1:
+    resolution: {integrity: sha512-F3PH5k5juxom4xktynS7MoFY+NUWH5LC4CnH11YB8NPew+HLpmBLCybSAEyb2F+4pRXhuhWqFesoQd6DAyc2hw==}
+    engines: {node: '>=18'}
+
  brace-expansion@1.1.12:
    resolution: {integrity: sha512-9T9UjW3r0UW5c1Q7GTwllptXwhvYmEzFhzMfZ9H7FQWt+uZePjZPjBP/W1ZEyZ1twGWom5/56TF4lPcqjnDHcg==}

@@ -10501,6 +10598,12 @@ packages:
    resolution: {integrity: sha512-yQbXgO/OSZVD2IsiLlro+7Hf6Q18EJrKSEsdoMzKePKXct3gvD8oLcOQdIzGupr5Fj+EDe8gO/lxc1BzfMpxvA==}
    engines: {node: '>=8'}

+  braintrust@1.0.2:
+    resolution: {integrity: sha512-Zm8WbD18FLv5HN38TyUKw659vJdZALQ5VVVWTncMxZgPtFXeWoC7H0tInw/1EQ1pGq50v7WZtlaFBOXhUaf8DQ==}
+    hasBin: true
+    peerDependencies:
+      zod: ^3.25.34
+
  browserslist@4.26.2:
    resolution: {integrity: sha512-ECFzp6uFOSB+dcZ5BK/IBaGWssbSYBHvuMeMt3MMFyhI0Z8SqGgEkBLARgpRH3hutIgPVsALcMwbDrJqPxQ65A==}
    engines: {node: ^6 || ^7 || ^8 || ^9 || ^10 || ^11 || ^12 || >=13.7}
@@ -10593,6 +10696,10 @@ packages:
    resolution: {integrity: sha512-xlx1yCK2Oc1APsPXDL2LdlNP6+uu8OCDdhOBSVT279M/S+y75O30C2VuD8T2ogdePBBl7PfPF4504tnLgX3zfw==}
    engines: {node: '>=14.16'}

+  camelcase@8.0.0:
+    resolution: {integrity: sha512-8WB3Jcas3swSvjIeA2yvCJ+Miyz5l1ZmB6HFb9R1317dt9LCQoswg/BGrmAmkWVEszSrrg4RwmO46qIm2OEnSA==}
+    engines: {node: '>=16'}
+
  camelize@1.0.1:
    resolution: {integrity: sha512-dU+Tx2fsypxTgtLoE36npi3UqcjSSMNYfkqgmoEhtZrraP5VWq0K7FkWVTYa8eMPtnU/G2txVsfdCJTn9uzpuQ==}

@@ -10682,6 +10789,9 @@ packages:
    resolution: {integrity: sha512-VqR8m68vM46BNnuZ5NtnGBKIE/DfN0cRIzg9n40EIq9NOv90ayxLBXA8fXC5gquFRGJSTRqBq25Jt2ECLR431Q==}
    engines: {node: '>= 6'}

+  cheminfo-types@1.8.1:
+    resolution: {integrity: sha512-FRcpVkox+cRovffgqNdDFQ1eUav+i/Vq/CUd1hcfEl2bevntFlzznL+jE8g4twl6ElB7gZjCko6pYpXyMn+6dA==}
+
  chevrotain-allstar@0.3.1:
    resolution: {integrity: sha512-b7g+y9A0v4mxCW1qUhf3BSVPg+/NvGErk/dOkrDaHA0nQIQGAtrOjlX//9OQtRlSCy+x9rfB5N8yC71lH1nvMw==}
    peerDependencies:
@@ -10754,10 +10864,18 @@ packages:
    resolution: {integrity: sha512-aCj4O5wKyszjMmDT4tZj93kxyydN/K5zPWSCe6/0AV/AA1pqe5ZBIw0a2ZfPQV7lL5/yb5HsUreJ6UFAF1tEQw==}
    engines: {node: '>=18'}

+  cli-progress@3.12.0:
+    resolution: {integrity: sha512-tRkV3HJ1ASwm19THiiLIXLO7Im7wlTuKnvkYaTkyoAPefqjNg7W7DHKUlGRxy9vxDvbyCYQkQozvptuMkGCg8A==}
+    engines: {node: '>=4'}
+
  cli-spinners@2.9.2:
    resolution: {integrity: sha512-ywqV+5MmyL4E7ybXgKys4DugZbX0FC6LnwrhjuykIjnK9k8OQacQ7axGKnjDXWNhns0xot3bZI5h55H8yo9cJg==}
    engines: {node: '>=6'}

+  cli-table3@0.6.5:
+    resolution: {integrity: sha512-+W/5efTR7y5HRD7gACw9yQjqMVvEMLBHmboM/kPWam+H+Hmyrgjh6YncVKK122YZkXrLudzTuAukUw9FnMf7IQ==}
+    engines: {node: 10.* || >= 12.*}
+
  cli-truncate@2.1.0:
    resolution: {integrity: sha512-n8fOixwDD6b/ObinzTrp1ZKFzbgvKZvuz/TvejnLn1aQfC6r52XEx85FmuC+3HI+JM7coBRXUvNqEU2PHVrHpg==}
    engines: {node: '>=8'}
@@ -10950,6 +11068,15 @@ packages:
    resolution: {integrity: sha512-6FqVXeETqWPoGcfzrXb37E50NP0LXT8kAMu5ooZayhWWdgEY4lBEEcbQNXtkuKQsGduxiIcI4gOTsxTmuq/bSg==}
    engines: {node: '>= 14'}

+  compute-cosine-similarity@1.1.0:
+    resolution: {integrity: sha512-FXhNx0ILLjGi9Z9+lglLzM12+0uoTnYkHm7GiadXDAr0HGVLm25OivUS1B/LPkbzzvlcXz/1EvWg9ZYyJSdhTw==}
+
+  compute-dot@1.1.0:
+    resolution: {integrity: sha512-L5Ocet4DdMrXboss13K59OK23GXjiSia7+7Ukc7q4Bl+RVpIXK2W9IHMbWDZkh+JUEvJAwOKRaJDiFUa1LTnJg==}
+
+  compute-l2norm@1.1.0:
+    resolution: {integrity: sha512-6EHh1Elj90eU28SXi+h2PLnTQvZmkkHWySpoFz+WOlVNLz3DQoC4ISUHSV9n5jMxPHtKGJ01F4uu2PsXBB8sSg==}
+
  compute-scroll-into-view@3.1.1:
    resolution: {integrity: sha512-VRhuHOLoKYOy4UbilLbUzbYg93XLjv2PncJC50EuTWPA3gaja1UjBsUP/D/9/juV3vQFr6XBEzn9KCAHdUvOHw==}

@@ -12363,6 +12490,10 @@ packages:
    resolution: {integrity: sha512-mQw+2fkQbALzQ7V0MY0IqdnXNOeTtP4r0lN9z7AAawCXgqea7bDii20AYrIBrFd/Hx0M2Ocz6S111CaFkUcb0Q==}
    engines: {node: '>=0.8.x'}

+  eventsource-parser@1.1.2:
+    resolution: {integrity: sha512-v0eOBUbiaFojBu2s2NPBfYUoRR9GjcDNvCXVaqEf5vVfpIAh9f8RCo4vXTP8c63QRKCFwoLpMpTdPwwhEKVgzA==}
+    engines: {node: '>=14.18'}
+
  eventsource-parser@3.0.3:
    resolution: {integrity: sha512-nVpZkTMM9rF6AQ9gPJpFsNAMt48wIzB5TQgiTLdHiuO8XEDhUgZEhqKlZWXbIzo9VmJ/HvysHqEaVeD5v9TPvA==}
    engines: {node: '>=20.0.0'}
@@ -12581,6 +12712,9 @@ packages:
  fflate@0.8.2:
    resolution: {integrity: sha512-cPJU47OaAoCbg0pBvzsgpTPhmhqI5eJjh/JIu8tPj5q+T7iLvW/JAYUqmE7KOB4R1ZyEhzBaIQpQpardBF5z8A==}

+  fft.js@4.0.4:
+    resolution: {integrity: sha512-f9c00hphOgeQTlDyavwTtu6RiK8AIFjD6+jvXkNkpeQ7rirK3uFWVpalkoS4LAwbdX7mfZ8aoBfFVQX1Re/8aw==}
+
  figures@3.2.0:
    resolution: {integrity: sha512-yaduQFRKLXYOGgEn6AZau90j3ggSOyiqXU0F9JZfeXYhNa+Jk4X+s45A2zg5jns87GAFa34BBm2kXw4XpNcbdg==}
    engines: {node: '>=8'}
@@ -13597,6 +13731,9 @@ packages:
  is-alphanumerical@2.0.1:
    resolution: {integrity: sha512-hmbYhX/9MUMF5uh7tOXyK/n0ZvWpad5caBA17GsC6vyuCqaWliRG5K1qS9inmUhEMaOBIW7/whAnSwveW/LtZw==}

+  is-any-array@2.0.1:
+    resolution: {integrity: sha512-UtilS7hLRu++wb/WBAw9bNuP1Eg04Ivn1vERJck8zJthEvXCBEBpGR/33u/xLKWEQf95803oalHrVDptcAvFdQ==}
+
  is-array-buffer@3.0.5:
    resolution: {integrity: sha512-DDfANUiiG2wC1qawP66qlTugJeL5HyzMpfr8lLK+jMQirGzNod0B12cFB/9q838Ru27sBwfw78/rdoU7RERz6A==}
    engines: {node: '>= 0.4'}
@@ -14247,6 +14384,9 @@ packages:
  libpg-query@15.2.0:
    resolution: {integrity: sha512-fBi14tsU3/ahtfU0C+/tvmtlt0EJM6kXe0Dl4/cSbiNmJ3IwLIuk3lawwiK5ByXU4sXwyAzbkvSl26jVXzYuaA==}

+  libpg-query@17.6.0:
+    resolution: {integrity: sha512-r4zOTcLTGYS5PlLQAicJ6Yi/tvZFag42YUuNEO8pi8bwt/ZZ4kj514J4QV5bOx0mZzPLF6agbfNXQVxGgmHR8g==}
+
  light-my-request@5.14.0:
    resolution: {integrity: sha512-aORPWntbpH5esaYpGOOmri0OHDOe3wC5M2MQxZ9dvMLZm6DnaAn0kJlcbU9hwsQgLzmZyReKwFwwPkR+nHu5kA==}

@@ -14258,6 +14398,9 @@ packages:
    resolution: {integrity: sha512-/vlFKAoH5Cgt3Ie+JLhRbwOsCQePABiU3tJ1egGvyQ+33R/vcwM2Zl2QR/LzjsBeItPt3oSVXapn+m4nQDvpzw==}
    engines: {node: '>=14'}

+  linear-sum-assignment@1.0.9:
+    resolution: {integrity: sha512-1T2Ek3sxpt2mBHeBFMRJEikiIK/yIOwf+mrxv/DkAU/5ddnCMndZL//hFH7QuHa1tbaQADzsf9t7rkGZKqoFfQ==}
+
  linebreak@1.1.0:
    resolution: {integrity: sha512-MHp03UImeVhB7XZtjd0E4n6+3xr5Dq/9xI/5FptGk5FrbDR3zagPa2DS6U8ks/3HjbKWG9Q1M2ufOzxV2qLYSQ==}

@@ -15119,6 +15262,24 @@ packages:
    engines: {node: '>=10'}
    hasBin: true

+  ml-array-max@1.2.4:
+    resolution: {integrity: sha512-BlEeg80jI0tW6WaPyGxf5Sa4sqvcyY6lbSn5Vcv44lp1I2GR6AWojfUvLnGTNsIXrZ8uqWmo8VcG1WpkI2ONMQ==}
+
+  ml-array-min@1.2.3:
+    resolution: {integrity: sha512-VcZ5f3VZ1iihtrGvgfh/q0XlMobG6GQ8FsNyQXD3T+IlstDv85g8kfV0xUG1QPRO/t21aukaJowDzMTc7j5V6Q==}
+
+  ml-array-rescale@1.3.7:
+    resolution: {integrity: sha512-48NGChTouvEo9KBctDfHC3udWnQKNKEWN0ziELvY3KG25GR5cA8K8wNVzracsqSW1QEkAXjTNx+ycgAv06/1mQ==}
+
+  ml-matrix@6.12.1:
+    resolution: {integrity: sha512-TJ+8eOFdp+INvzR4zAuwBQJznDUfktMtOB6g/hUcGh3rcyjxbz4Te57Pgri8Q9bhSQ7Zys4IYOGhFdnlgeB6Lw==}
+
+  ml-spectra-processing@14.18.2:
+    resolution: {integrity: sha512-n9aXXxdf+ogpYeU6FLZgClOteAE10DyUMkjeBGnvew9vmewKLfINVbMZBosW4Q1aqFDMsZjaItiJJ/UCApTIKg==}
+
+  ml-xsadd@3.0.1:
+    resolution: {integrity: sha512-Fz2q6dwgzGM8wYKGArTUTZDGa4lQFA2Vi6orjGeTVRy22ZnQFKlJuwS9n8NRviqz1KHAHAzdKJwbnYhdo38uYg==}
+
  mlly@1.8.0:
    resolution: {integrity: sha512-l8D9ODSRWLe2KHJSifWGwBqpTZXIXTeo8mlKjY+E2HAakaTeNpqAyBZ8GSqLzHgw4XmHmC8whvpjJNMbFZN7/g==}

@@ -15199,6 +15360,10 @@ packages:
  muggle-string@0.4.1:
    resolution: {integrity: sha512-VNTrAak/KhO2i8dqqnqnAHOa3cYBwXEZe9h+D5h/1ZqFSTEFHdM65lR7RoIqq3tBBYavsOXV84NoHXZ0AkPyqQ==}

+  mustache@4.2.0:
+    resolution: {integrity: sha512-71ippSywq5Yb7/tVYyGbkBggbU8H3u5Rz56fH60jGFgr8uHwxs+aSKeqmluIVzM0m0kB7xQjKS6qPfd0b2ZoqQ==}
+    hasBin: true
+
  mute-stream@0.0.8:
    resolution: {integrity: sha512-nnbWWOkoWyUsTjKrhgD0dcz22mdkSnpYqbEjIm2nhwhuxlSkpywJmBo8h0ZqJdkp73mb90SssHkN4rsRaBAfAA==}

@@ -18212,8 +18377,8 @@ packages:
    resolution: {integrity: sha512-TDQg6Nh7e2DaLmMjqpJt7Wwh2Au5loVPH2gwKMkD89e3pb5F/zMQGWt22asEXme031jEmhjazixFH0lwMgbRJA==}
    engines: {node: '>=18.0.0'}

-  supabase@2.58.5:
-    resolution: {integrity: sha512-mYZSkUIePTdmwlHd26Pff8wpmjfre8gcuWzrc5QqhZgZvCXugVzAQQhcjaQisw5kusbPQWNIjUwcHYEKmejhPw==}
+  supabase@2.67.1:
+    resolution: {integrity: sha512-d/trGytTjB/Hi625zHh2RFFttjLOkSdRTsY/N1pjxoAfG0blDiHKqPwu12VJYitg4nzgjfhjnD3pLyfU0ko5vg==}
    engines: {npm: '>=8'}
    hasBin: true

@@ -18316,6 +18481,10 @@ packages:
  tdigest@0.1.2:
    resolution: {integrity: sha512-+G0LLgjjo9BZX2MfdvPfH+MKLCrxlXSYec5DaPYP1fe6Iyhf0/fSmJ0bFiZ1F8BT6cGXl2LpltQptzjXKWEkKA==}

+  termi-link@1.1.0:
+    resolution: {integrity: sha512-2qSN6TnomHgVLtk+htSWbaYs4Rd2MH/RU7VpHTy6MBstyNyWbM4yKd1DCYpE3fDg8dmGWojXCngNi/MHCzGuAA==}
+    engines: {node: '>=12'}
+
  terser-webpack-plugin@5.3.14:
    resolution: {integrity: sha512-vkZjpUjb6OMS7dhV+tILUW6BhpDR7P2L/aQSAv+Uwk+m8KATX9EccViHTJR2qDtACKPIYndLGCyl3FMo+r2LMw==}
    engines: {node: '>= 10.13.0'}
@@ -19158,6 +19327,12 @@ packages:
    resolution: {integrity: sha512-OljLrQ9SQdOUqTaQxqL5dEfZWrXExyyWsozYlAWFawPVNuD83igl7uJD2RTkNMbniIYgt8l81eCJGIdQF7avLQ==}
    engines: {node: ^14.17.0 || ^16.13.0 || >=18.0.0}

+  validate.io-array@1.0.6:
+    resolution: {integrity: sha512-DeOy7CnPEziggrOO5CZhVKJw6S3Yi7e9e65R1Nl/RTN1vTQKnzjfvks0/8kQ40FP/dsjRAOd4hxmJ7uLa6vxkg==}
+
+  validate.io-function@1.0.2:
+    resolution: {integrity: sha512-LlFybRJEriSuBnUhQyG5bwglhh50EpTL2ul23MPIuR1odjO7XaMLFV8vHGwp7AZciFxtYOeiSCT5st+XSPONiQ==}
+
  valtio@1.12.0:
    resolution: {integrity: sha512-co8NkCHeY0NsL0XsL/cSICt5VhTjwZlYT8mi50dYY5thx3r3w1D15A04Lvs9WL/y/Rf98vUKY5PAAJCTLHvkJw==}
    engines: {node: '>=12.20.0'}
@@ -19543,6 +19718,10 @@ packages:
    resolution: {integrity: sha512-o0cyEG0e8GPzT4iGHphIOh0cJOV8fivsXxddQasHPHfoZf1ZexrfeA21w2NaEN1RHE+fXlfISmOE8R9N3u3Qig==}
    engines: {node: '>=12'}

+  widest-line@5.0.0:
+    resolution: {integrity: sha512-c9bZp7b5YtRj2wOe6dlj32MK+Bx/M/d+9VB2SHM1OtsUHR0aV0tdP6DWh/iMt0kWi1t5g1Iudu6hQRNd1A4PVA==}
+    engines: {node: '>=18'}
+
  wordwrap@1.0.0:
    resolution: {integrity: sha512-gvVzJFlPycKc5dZN4yPkP8w7Dc37BtP1yczEneOb4uq34pXZcvrtRTmWV8W+Ume+XCxKgbjM+nevkyFPMybd4Q==}

@@ -19558,6 +19737,10 @@ packages:
    resolution: {integrity: sha512-si7QWI6zUMq56bESFvagtmzMdGOtoxfR+Sez11Mobfc7tm+VkUckk9bW2UeffTGVUbOksxmSw0AA2gs8g71NCQ==}
    engines: {node: '>=12'}

+  wrap-ansi@9.0.2:
+    resolution: {integrity: sha512-42AtmgqjV+X1VpdOfyTGOYRi0/zsoLqtXQckTmqTeybT+BDIbM/Guxo7x3pE2vtpr1ok6xRqM9OpBe+Jyoqyww==}
+    engines: {node: '>=18'}
+
  wrappy@1.0.2:
    resolution: {integrity: sha512-l4Sp/DRseor9wL6EvV2+TuQn63dMkPjZ/sp9XkghTEbV9KlPS1xUsZ3u7/IQO4wxtcFB4bgpQPRcR3QCvezPcQ==}

@@ -19797,6 +19980,10 @@ snapshots:
      eventsource-parser: 3.0.6
      zod: 3.25.76

+  '@ai-sdk/provider@1.1.3':
+    dependencies:
+      json-schema: 0.4.0
+
  '@ai-sdk/provider@2.0.0':
    dependencies:
      json-schema: 0.4.0
@@ -21204,6 +21391,9 @@ snapshots:
      react-dom: 18.3.1(react@18.3.1)
      react-is: 17.0.2

+  '@colors/colors@1.5.0':
+    optional: true
+
  '@contentlayer2/cli@0.4.3(esbuild@0.25.2)(markdown-wasm@1.2.0)(supports-color@8.1.1)':
    dependencies:
      '@contentlayer2/core': 0.4.3(esbuild@0.25.2)(markdown-wasm@1.2.0)(supports-color@8.1.1)
@@ -23329,6 +23519,8 @@ snapshots:
      - bufferutil
      - utf-8-validate

+  '@next/env@14.2.35': {}
+
  '@next/env@15.5.9': {}

  '@next/env@16.0.10': {}
@@ -24823,6 +25015,8 @@ snapshots:

  '@pgsql/types@15.0.2': {}

+  '@pgsql/types@17.6.2': {}
+
  '@pkgjs/parseargs@0.11.0':
    optional: true

@@ -29322,6 +29516,10 @@ snapshots:
      lz-string: 1.5.0
      uuid: 9.0.1

+  '@vercel/functions@1.6.0(@aws-sdk/credential-provider-web-identity@3.830.0)':
+    optionalDependencies:
+      '@aws-sdk/credential-provider-web-identity': 3.830.0
+
  '@vercel/functions@2.1.0(@aws-sdk/credential-provider-web-identity@3.830.0)':
    optionalDependencies:
      '@aws-sdk/credential-provider-web-identity': 3.830.0
@@ -30098,6 +30296,21 @@ snapshots:

  auto-bind@4.0.0: {}

+  autoevals@0.0.131(encoding@0.1.13)(ws@8.18.3):
+    dependencies:
+      ajv: 8.17.1
+      compute-cosine-similarity: 1.1.0
+      js-levenshtein: 1.1.6
+      js-yaml: 4.1.1
+      linear-sum-assignment: 1.0.9
+      mustache: 4.2.0
+      openai: 4.104.0(encoding@0.1.13)(ws@8.18.3)(zod@3.25.76)
+      zod: 3.25.76
+      zod-to-json-schema: 3.25.0(zod@3.25.76)
+    transitivePeerDependencies:
+      - encoding
+      - ws
+
  autolinker@0.28.1:
    dependencies:
      gulp-header: 1.8.12
@@ -30205,6 +30418,8 @@ snapshots:

  binary-extensions@2.2.0: {}

+  binary-search@1.3.6: {}
+
  bindings@1.5.0:
    dependencies:
      file-uri-to-path: 1.0.0
@@ -30267,6 +30482,17 @@ snapshots:
      widest-line: 4.0.1
      wrap-ansi: 8.1.0

+  boxen@8.0.1:
+    dependencies:
+      ansi-align: 3.0.1
+      camelcase: 8.0.0
+      chalk: 5.4.1
+      cli-boxes: 3.0.0
+      string-width: 7.2.0
+      type-fest: 4.30.0
+      widest-line: 5.0.0
+      wrap-ansi: 9.0.2
+
  brace-expansion@1.1.12:
    dependencies:
      balanced-match: 1.0.2
@@ -30280,6 +30506,37 @@ snapshots:
    dependencies:
      fill-range: 7.1.1

+  braintrust@1.0.2(@aws-sdk/credential-provider-web-identity@3.830.0)(supports-color@8.1.1)(zod@3.25.76):
+    dependencies:
+      '@ai-sdk/provider': 1.1.3
+      '@next/env': 14.2.35
+      '@vercel/functions': 1.6.0(@aws-sdk/credential-provider-web-identity@3.830.0)
+      argparse: 2.0.1
+      boxen: 8.0.1
+      chalk: 4.1.2
+      cli-progress: 3.12.0
+      cli-table3: 0.6.5
+      cors: 2.8.5
+      dotenv: 16.5.0
+      esbuild: 0.25.2
+      eventsource-parser: 1.1.2
+      express: 4.22.1(supports-color@8.1.1)
+      graceful-fs: 4.2.11
+      http-errors: 2.0.1
+      minimatch: 9.0.5
+      mustache: 4.2.0
+      pluralize: 8.0.0
+      simple-git: 3.28.0(supports-color@8.1.1)
+      slugify: 1.6.6
+      source-map: 0.7.6
+      termi-link: 1.1.0
+      uuid: 9.0.1
+      zod: 3.25.76
+      zod-to-json-schema: 3.25.0(zod@3.25.76)
+    transitivePeerDependencies:
+      - '@aws-sdk/credential-provider-web-identity'
+      - supports-color
+
  browserslist@4.26.2:
    dependencies:
      baseline-browser-mapping: 2.8.5
@@ -30416,6 +30673,8 @@ snapshots:

  camelcase@7.0.1: {}

+  camelcase@8.0.0: {}
+
  camelize@1.0.1: {}

  caniuse-api@3.0.0:
@@ -30534,6 +30793,8 @@ snapshots:
      parse5: 7.1.2
      parse5-htmlparser2-tree-adapter: 7.0.0

+  cheminfo-types@1.8.1: {}
+
  chevrotain-allstar@0.3.1(chevrotain@11.0.3):
    dependencies:
      chevrotain: 11.0.3
@@ -30606,8 +30867,18 @@ snapshots:
    dependencies:
      restore-cursor: 5.1.0

+  cli-progress@3.12.0:
+    dependencies:
+      string-width: 4.2.3
+
  cli-spinners@2.9.2: {}

+  cli-table3@0.6.5:
+    dependencies:
+      string-width: 4.2.3
+    optionalDependencies:
+      '@colors/colors': 1.5.0
+
  cli-truncate@2.1.0:
    dependencies:
      slice-ansi: 3.0.0
@@ -30784,6 +31055,23 @@ snapshots:
      normalize-path: 3.0.0
      readable-stream: 4.6.0

+  compute-cosine-similarity@1.1.0:
+    dependencies:
+      compute-dot: 1.1.0
+      compute-l2norm: 1.1.0
+      validate.io-array: 1.0.6
+      validate.io-function: 1.0.2
+
+  compute-dot@1.1.0:
+    dependencies:
+      validate.io-array: 1.0.6
+      validate.io-function: 1.0.2
+
+  compute-l2norm@1.1.0:
+    dependencies:
+      validate.io-array: 1.0.6
+      validate.io-function: 1.0.2
+
  compute-scroll-into-view@3.1.1: {}

  concat-map@0.0.1: {}
@@ -32263,6 +32551,8 @@ snapshots:

  events@3.3.0: {}

+  eventsource-parser@1.1.2: {}
+
  eventsource-parser@3.0.3: {}

  eventsource-parser@3.0.6: {}
@@ -32582,6 +32872,8 @@ snapshots:

  fflate@0.8.2: {}

+  fft.js@4.0.4: {}
+
  figures@3.2.0:
    dependencies:
      escape-string-regexp: 1.0.5
@@ -33836,6 +34128,8 @@ snapshots:
      is-alphabetical: 2.0.1
      is-decimal: 2.0.1

+  is-any-array@2.0.1: {}
+
  is-array-buffer@3.0.5:
    dependencies:
      call-bind: 1.0.8
@@ -34448,6 +34742,10 @@ snapshots:
      - encoding
      - supports-color

+  libpg-query@17.6.0:
+    dependencies:
+      '@pgsql/types': 17.6.2
+
  light-my-request@5.14.0:
    dependencies:
      cookie: 0.7.2
@@ -34458,6 +34756,12 @@ snapshots:

  lilconfig@3.1.3: {}

+  linear-sum-assignment@1.0.9:
+    dependencies:
+      cheminfo-types: 1.8.1
+      ml-matrix: 6.12.1
+      ml-spectra-processing: 14.18.2
+
  linebreak@1.1.0:
    dependencies:
      base64-js: 0.0.8
@@ -35975,6 +36279,36 @@ snapshots:

  mkdirp@3.0.1: {}

+  ml-array-max@1.2.4:
+    dependencies:
+      is-any-array: 2.0.1
+
+  ml-array-min@1.2.3:
+    dependencies:
+      is-any-array: 2.0.1
+
+  ml-array-rescale@1.3.7:
+    dependencies:
+      is-any-array: 2.0.1
+      ml-array-max: 1.2.4
+      ml-array-min: 1.2.3
+
+  ml-matrix@6.12.1:
+    dependencies:
+      is-any-array: 2.0.1
+      ml-array-rescale: 1.3.7
+
+  ml-spectra-processing@14.18.2:
+    dependencies:
+      binary-search: 1.3.6
+      cheminfo-types: 1.8.1
+      fft.js: 4.0.4
+      is-any-array: 2.0.1
+      ml-matrix: 6.12.1
+      ml-xsadd: 3.0.1
+
+  ml-xsadd@3.0.1: {}
+
  mlly@1.8.0:
    dependencies:
      acorn: 8.15.0
@@ -36054,6 +36388,8 @@ snapshots:

  muggle-string@0.4.1: {}

+  mustache@4.2.0: {}
+
  mute-stream@0.0.8: {}

  mute-stream@2.0.0: {}
@@ -36520,7 +36856,7 @@ snapshots:

  number-flow@0.3.7: {}

-  nuqs@1.19.1(next@15.5.9(@opentelemetry/api@1.9.0)(@playwright/test@1.56.1)(react-dom@18.3.1(react@18.3.1))(react@18.3.1)(sass@1.77.4)):
+  nuqs@1.19.1(next@15.5.9(@babel/core@7.28.4(supports-color@8.1.1))(@opentelemetry/api@1.9.0)(@playwright/test@1.56.1)(react-dom@18.3.1(react@18.3.1))(react@18.3.1)(sass@1.77.4)):
    dependencies:
      mitt: 3.0.1
      next: 15.5.9(@babel/core@7.28.4(supports-color@8.1.1))(@opentelemetry/api@1.9.0)(@playwright/test@1.56.1)(babel-plugin-macros@3.1.0)(react-dom@18.3.1(react@18.3.1))(react@18.3.1)(sass@1.77.4)
@@ -39921,7 +40257,7 @@ snapshots:
    dependencies:
      openapi-fetch: 0.6.2

-  supabase@2.58.5(supports-color@8.1.1):
+  supabase@2.67.1(supports-color@8.1.1):
    dependencies:
      bin-links: 6.0.0
      https-proxy-agent: 7.0.6(supports-color@8.1.1)
@@ -40065,6 +40401,8 @@ snapshots:
    dependencies:
      bintrees: 1.0.2

+  termi-link@1.1.0: {}
+
  terser-webpack-plugin@5.3.14(esbuild@0.25.2)(webpack@5.94.0(esbuild@0.25.2)):
    dependencies:
      '@jridgewell/trace-mapping': 0.3.31
@@ -40898,6 +41236,10 @@ snapshots:

  validate-npm-package-name@5.0.1: {}

+  validate.io-array@1.0.6: {}
+
+  validate.io-function@1.0.2: {}
+
  valtio@1.12.0(@types/react@18.3.3)(react@18.3.1):
    dependencies:
      derive-valtio: 0.1.0(valtio@1.12.0(@types/react@18.3.3)(react@18.3.1))
@@ -41471,6 +41813,10 @@ snapshots:
    dependencies:
      string-width: 5.1.2

+  widest-line@5.0.0:
+    dependencies:
+      string-width: 7.2.0
+
  wordwrap@1.0.0: {}

  wrap-ansi@6.2.0:
@@ -41491,6 +41837,12 @@ snapshots:
      string-width: 5.1.2
      strip-ansi: 7.1.0

+  wrap-ansi@9.0.2:
+    dependencies:
+      ansi-styles: 6.2.1
+      string-width: 7.2.0
+      strip-ansi: 7.1.0
+
  wrappy@1.0.2: {}

  write-file-atomic@7.0.0:
--- a/pnpm-workspace.yaml
+++ b/pnpm-workspace.yaml
@@ -51,6 +51,7 @@ minimumReleaseAgeExclude:
  - iceberg-js
  - '@vitejs/plugin-rsc'
  - stripe-experiment-sync # TODO(matlin) remove, temp just to unblock launch
+  - braintrust

 onlyBuiltDependencies:
  - supabase
--- a/turbo.json
+++ b/turbo.json
@@ -104,6 +104,8 @@
        "DEFAULT_PROJECT_NAME",
        "DEFAULT_ORGANIZATION_NAME",
        "OPENAI_API_KEY",
+        "BRAINTRUST_API_KEY",
+        "BRAINTRUST_PROJECT_ID",
        "AUTH_JWT_SECRET",
        "LOGFLARE_API_KEY",
        "LOGFLARE_PUBLIC_ACCESS_TOKEN",