fix(errors): show actual host in 404 message instead of Ollama hint (#926)

When an OpenAI-compatible provider returns a 404, the user-facing error
message hardcoded "for Ollama: http://127.0.0.1:11434/v1" as a hint
regardless of the configured base URL. Users on remote providers
(NVIDIA NIM, OpenRouter, etc.) read this as the app ignoring their
custom OPENAI_BASE_URL and routing to localhost.

Plumb the request URL through the classifier and marker so the
user-facing message can name the actual host. Localhost endpoints keep
the existing Ollama-flavored guidance for backward compatibility.

- classifyOpenAIHttpFailure now accepts an optional url and produces a
  host-aware hint for non-localhost 404s
- the [openai_category=...] marker carries an optional host segment
- mapOpenAICompatibilityFailureToAssistantMessage branches on host to
  show "Endpoint at <host> returned 404. Verify OPENAI_BASE_URL is
  correct and the selected model (<model>) is supported by this
  provider." for remote URLs
- backward compatibility preserved when no URL is available

bun.lock

@@ -28,7 +28,6 @@
         "@opentelemetry/sdk-trace-base": "2.6.1",
         "@opentelemetry/sdk-trace-node": "2.6.1",
         "@opentelemetry/semantic-conventions": "1.40.0",
-        "@vscode/ripgrep": "^1.17.1",
         "ajv": "8.18.0",
         "auto-bind": "5.0.1",
         "axios": "1.15.0",
@@ -462,8 +461,6 @@
 
     "@types/react": ["@types/react@19.2.14", "", { "dependencies": { "csstype": "^3.2.2" } }, "sha512-ilcTH/UniCkMdtexkoCN0bI7pMcJDvmQFPvuPvmEaYA/NSfFTAgdUSLAoVjaRJm7+6PvcM+q1zYOwS4wTYMF9w=="],
 
-    "@vscode/ripgrep": ["@vscode/ripgrep@1.17.1", "", { "dependencies": { "https-proxy-agent": "^7.0.2", "proxy-from-env": "^1.1.0", "yauzl": "^2.9.2" } }, "sha512-xTs7DGyAO3IsJYOCTBP8LnTvPiYVKEuyv8s0xyJDBXfs8rhBfqnZPvb6xDT+RnwWzcXqW27xLS/aGrkjX7lNWw=="],
-
     "accepts": ["accepts@2.0.0", "", { "dependencies": { "mime-types": "^3.0.0", "negotiator": "^1.0.0" } }, "sha512-5cvg6CtKwfgdmVqY1WIiXKc3Q1bkRqGLi+2W/6ao+6Y7gu/RCwRuAhGEzh5B4KlszSuTLgZYuqFqo5bImjNKng=="],
 
     "agent-base": ["agent-base@7.1.4", "", {}, "sha512-MnA+YT8fwfJPgBx3m60MNqakm30XOkyIoH1y6huTQvC0PwZG7ki8NacLBcrPbNoo8vEZy7Jpuk7+jMO+CUovTQ=="],
@@ -494,8 +491,6 @@
 
     "bowser": ["bowser@2.14.1", "", {}, "sha512-tzPjzCxygAKWFOJP011oxFHs57HzIhOEracIgAePE4pqB3LikALKnSzUyU4MGs9/iCEUuHlAJTjTc5M+u7YEGg=="],
 
-    "buffer-crc32": ["buffer-crc32@0.2.13", "", {}, "sha512-VO9Ht/+p3SN7SKWqcrgEzjGbRSJYTx+Q1pTQC0wrWqHx0vpJraQ6GtHx8tvcg1rlK1byhU5gccxgOgj7B0TDkQ=="],
-
     "buffer-equal-constant-time": ["buffer-equal-constant-time@1.0.1", "", {}, "sha512-zRpUiDwd/xk6ADqPMATG8vc9VPrkck7T07OIx0gnjmJAnHnTVXNQG3vfvWNuiZIkwu9KrKdA1iJKfsfTVxE6NA=="],
 
     "bun-types": ["bun-types@1.3.11", "", { "dependencies": { "@types/node": "*" } }, "sha512-1KGPpoxQWl9f6wcZh57LvrPIInQMn2TQ7jsgxqpRzg+l0QPOFvJVH7HmvHo/AiPgwXy+/Thf6Ov3EdVn1vOabg=="],
@@ -614,8 +609,6 @@
 
     "fast-xml-parser": ["fast-xml-parser@5.5.8", "", { "dependencies": { "fast-xml-builder": "^1.1.4", "path-expression-matcher": "^1.2.0", "strnum": "^2.2.0" }, "bin": { "fxparser": "src/cli/cli.js" } }, "sha512-Z7Fh2nVQSb2d+poDViM063ix2ZGt9jmY1nWhPfHBOK2Hgnb/OW3P4Et3P/81SEej0J7QbWtJqxO05h8QYfK7LQ=="],
 
-    "fd-slicer": ["fd-slicer@1.1.0", "", { "dependencies": { "pend": "~1.2.0" } }, "sha512-cE1qsB/VwyQozZ+q1dGxR8LBYNZeofhEdUNGSMbQD3Gw2lAzX9Zb3uIU6Ebc/Fmyjo9AWWfnn0AUCHqtevs/8g=="],
-
     "fflate": ["fflate@0.8.2", "", {}, "sha512-cPJU47OaAoCbg0pBvzsgpTPhmhqI5eJjh/JIu8tPj5q+T7iLvW/JAYUqmE7KOB4R1ZyEhzBaIQpQpardBF5z8A=="],
 
     "figures": ["figures@6.1.0", "", { "dependencies": { "is-unicode-supported": "^2.0.0" } }, "sha512-d+l3qxjSesT4V7v2fh+QnmFnUWv9lSpjarhShNTgBOfA0ttejbQUAlHLitbjkoRiDulW0OPoQPYIGhIC8ohejg=="],
@@ -794,8 +787,6 @@
 
     "path-to-regexp": ["path-to-regexp@8.4.1", "", {}, "sha512-fvU78fIjZ+SBM9YwCknCvKOUKkLVqtWDVctl0s7xIqfmfb38t2TT4ZU2gHm+Z8xGwgW+QWEU3oQSAzIbo89Ggw=="],
 
-    "pend": ["pend@1.2.0", "", {}, "sha512-F3asv42UuXchdzt+xXqfW1OGlVBe+mxa2mqI0pg5yAHZPvFmY3Y6drSf/GQ1A86WgWEN9Kzh/WrgKa6iGcHXLg=="],
-
     "picomatch": ["picomatch@4.0.4", "", {}, "sha512-QP88BAKvMam/3NxH6vj2o21R6MjxZUAd6nlwAS/pnGvN9IVLocLHxGYIzFhg6fUQ+5th6P4dv4eW9jX3DSIj7A=="],
 
     "pkce-challenge": ["pkce-challenge@5.0.1", "", {}, "sha512-wQ0b/W4Fr01qtpHlqSqspcj3EhBvimsdh0KlHhH8HRZnMsEa0ea2fTULOXOS9ccQr3om+GcGRk4e+isrZWV8qQ=="],
@@ -810,7 +801,7 @@
 
     "proxy-addr": ["proxy-addr@2.0.7", "", { "dependencies": { "forwarded": "0.2.0", "ipaddr.js": "1.9.1" } }, "sha512-llQsMLSUDUPT44jdrU/O37qlnifitDP+ZwrmmZcoSKyLKvtZxpyV0n2/bD/N4tBAAZ/gJEdZU7KMraoK1+XYAg=="],
 
-    "proxy-from-env": ["proxy-from-env@1.1.0", "", {}, "sha512-D+zkORCbA9f1tdWRK0RaCR3GPv50cMxcrz4X8k5LTSUD1Dkw47mKJEZQNunItRTkWwgtaUSo1RVFRIG9ZXiFYg=="],
+    "proxy-from-env": ["proxy-from-env@2.1.0", "", {}, "sha512-cJ+oHTW1VAEa8cJslgmUZrc+sjRKgAKl3Zyse6+PV38hZe/V6Z14TbCuXcan9F9ghlz4QrFr2c92TNF82UkYHA=="],
 
     "qrcode": ["qrcode@1.5.4", "", { "dependencies": { "dijkstrajs": "^1.0.1", "pngjs": "^5.0.0", "yargs": "^15.3.1" }, "bin": { "qrcode": "bin/qrcode" } }, "sha512-1ca71Zgiu6ORjHqFBDpnSMTR2ReToX4l1Au1VFLyVeBTFavzQnv5JxMFr3ukHVKpSrSA2MCk0lNJSykjUfz7Zg=="],
 
@@ -962,8 +953,6 @@
 
     "yargs-parser": ["yargs-parser@21.1.1", "", {}, "sha512-tVpsJW7DdjecAiFpbIB1e3qxIQsE6NoPc5/eTdrbbIC4h0LVsWhnoa3g+m2HclBIujHzsxZ4VJVA+GUuc2/LBw=="],
 
-    "yauzl": ["yauzl@2.10.0", "", { "dependencies": { "buffer-crc32": "~0.2.3", "fd-slicer": "~1.1.0" } }, "sha512-p4a9I6X6nu6IhoGmBqAcbJy1mlC4j27vEPZX9F4L4/vZT3Lyq1VkFHw/V/PUcB9Buo+DG3iHkT0x3Qya58zc3g=="],
-
     "yoctocolors": ["yoctocolors@2.1.2", "", {}, "sha512-CzhO+pFNo8ajLM2d2IW/R93ipy99LWjtwblvC1RsoSUMZgyLbYFr221TnSNT7GjGdYui6P459mw9JH/g/zW2ug=="],
 
     "zod": ["zod@3.25.76", "", {}, "sha512-gzUt/qt81nXsFGKIFcC3YnfEAx5NkunCfnDlvuBSSFS02bcXu4Lmea0AFIUwbLWxWPx3d9p8S5QoaujKcNQxcQ=="],
@@ -1380,8 +1369,6 @@
 
     "@smithy/uuid/tslib": ["tslib@2.8.1", "", {}, "sha512-oJFu94HQb+KVduSUQL7wnpmqnfmLsOA/nAh6b6EH0wCEoK0/mPeXU6c3wKDV83MkOuHPRHtSXKKU99IBazS/2w=="],
 
-    "axios/proxy-from-env": ["proxy-from-env@2.1.0", "", {}, "sha512-cJ+oHTW1VAEa8cJslgmUZrc+sjRKgAKl3Zyse6+PV38hZe/V6Z14TbCuXcan9F9ghlz4QrFr2c92TNF82UkYHA=="],
-
     "cli-highlight/chalk": ["chalk@4.1.2", "", { "dependencies": { "ansi-styles": "^4.1.0", "supports-color": "^7.1.0" } }, "sha512-oKnbhFyRIXpUuez8iBMmyEa4nbj4IOQyuhc/wy9kY7/WVPcwIO9VA668Pu8RkO7+0G76SLROeyw9CpQ061i4mA=="],
 
     "cli-highlight/yargs": ["yargs@16.2.0", "", { "dependencies": { "cliui": "^7.0.2", "escalade": "^3.1.1", "get-caller-file": "^2.0.5", "require-directory": "^2.1.1", "string-width": "^4.2.0", "y18n": "^5.0.5", "yargs-parser": "^20.2.2" } }, "sha512-D1mvvtDG0L5ft/jGWkLpG1+m0eQxOfaBvTNELraWj22wSVUMWxZUvYgJYcKh6jGGIkJFhH4IZPQhR4TKpc8mBw=="],
@@ -1442,8 +1429,6 @@
 
     "@aws-sdk/nested-clients/@smithy/util-base64/@smithy/util-buffer-from": ["@smithy/util-buffer-from@4.2.2", "", { "dependencies": { "@smithy/is-array-buffer": "^4.2.2", "tslib": "^2.6.2" } }, "sha512-FDXD7cvUoFWwN6vtQfEta540Y/YBe5JneK3SoZg9bThSoOAC/eGeYEua6RkBgKjGa/sz6Y+DuBZj3+YEY21y4Q=="],
 
-    "@mendable/firecrawl-js/axios/proxy-from-env": ["proxy-from-env@2.1.0", "", {}, "sha512-cJ+oHTW1VAEa8cJslgmUZrc+sjRKgAKl3Zyse6+PV38hZe/V6Z14TbCuXcan9F9ghlz4QrFr2c92TNF82UkYHA=="],
-
     "@opentelemetry/exporter-trace-otlp-grpc/@opentelemetry/core/@opentelemetry/semantic-conventions": ["@opentelemetry/semantic-conventions@1.28.0", "", {}, "sha512-lp4qAiMTD4sNWW4DbKLBkfiMZ4jbAboJIGOQr5DvciMRI494OapieI9qiODpOt0XBr1LjIDy1xAGAnVs5supTA=="],
 
     "@opentelemetry/exporter-trace-otlp-grpc/@opentelemetry/otlp-transformer/@opentelemetry/api-logs": ["@opentelemetry/api-logs@0.57.2", "", { "dependencies": { "@opentelemetry/api": "^1.3.0" } }, "sha512-uIX52NnTM0iBh84MShlpouI7UKqkZ7MrUszTmaypHBu4r7NofznSnQRfJ+uUeDtQDj6w8eFGg5KBLDAwAPz1+A=="],
@@ -1524,8 +1509,6 @@
 
     "cliui/wrap-ansi/ansi-styles": ["ansi-styles@4.3.0", "", { "dependencies": { "color-convert": "^2.0.1" } }, "sha512-zbB9rCJAT1rbjiVDb2hqKFHNYLxgtk8NURxZ3IZwD3F6NtxbXZQCnnSi1Lkx+IDohdPlFp222wVALIheZJQSEg=="],
 
-    "firecrawl/axios/proxy-from-env": ["proxy-from-env@2.1.0", "", {}, "sha512-cJ+oHTW1VAEa8cJslgmUZrc+sjRKgAKl3Zyse6+PV38hZe/V6Z14TbCuXcan9F9ghlz4QrFr2c92TNF82UkYHA=="],
-
     "form-data/mime-types/mime-db": ["mime-db@1.52.0", "", {}, "sha512-sPU4uV7dYlvtWJxwwxHD0PuihVNiE7TyAbQ5SWxDCB9mUYvOgroQOwYQQOKPJ8CIbE+1ETVlOoK1UC2nU3gYvg=="],
 
     "qrcode/yargs/cliui": ["cliui@6.0.0", "", { "dependencies": { "string-width": "^4.2.0", "strip-ansi": "^6.0.0", "wrap-ansi": "^6.2.0" } }, "sha512-t6wbgtoCXvAzst7QgXxJYqPt0usEfbgQdftEPbLL/cvv6HPE5VgvqCuAIDR0NgU52ds6rFwqrgakNLrHEjCbrQ=="],

package.json

@@ -74,7 +74,6 @@
     "@opentelemetry/sdk-trace-base": "2.6.1",
     "@opentelemetry/sdk-trace-node": "2.6.1",
     "@opentelemetry/semantic-conventions": "1.40.0",
-    "@vscode/ripgrep": "^1.17.1",
     "ajv": "8.18.0",
     "auto-bind": "5.0.1",
     "axios": "1.15.0",

scripts/build.ts

@@ -472,11 +472,6 @@ ${exports}
     '@aws-sdk/credential-providers',
     '@azure/identity',
     'google-auth-library',
-    // @vscode/ripgrep ships a platform-specific binary alongside its
-    // index.js and resolves the path via __dirname at runtime. Bundling
-    // would freeze the build host's absolute path into dist/cli.mjs, so we
-    // keep it external and rely on the npm package being installed.
-    '@vscode/ripgrep',
   ],
 })
 

src/services/api/errors.openaiCompatibility.test.ts

@@ -28,6 +28,38 @@ test('maps endpoint_not_found category markers to actionable setup guidance', ()
   expect(text).toContain('/v1')
 })
 
+test('endpoint_not_found from a remote host shows the actual host, not Ollama (issue #926)', () => {
+  const error = APIError.generate(
+    404,
+    undefined,
+    'OpenAI API error 404: Not Found [openai_category=endpoint_not_found,host=integrate.api.nvidia.com] Hint: Endpoint at integrate.api.nvidia.com returned 404.',
+    new Headers(),
+  )
+
+  const message = getAssistantMessageFromError(error, 'moonshotai/kimi-k2.5-thinking')
+  const text = getFirstText(message)
+
+  expect(text).toContain('integrate.api.nvidia.com')
+  expect(text).toContain('moonshotai/kimi-k2.5-thinking')
+  expect(text).not.toContain('Ollama')
+  expect(text).not.toContain('11434')
+})
+
+test('endpoint_not_found without a host falls back to the Ollama-aware message', () => {
+  const error = APIError.generate(
+    404,
+    undefined,
+    'OpenAI API error 404: Not Found [openai_category=endpoint_not_found] Hint: Confirm OPENAI_BASE_URL includes /v1.',
+    new Headers(),
+  )
+
+  const message = getAssistantMessageFromError(error, 'qwen2.5-coder:7b')
+  const text = getFirstText(message)
+
+  expect(text).toContain('Provider endpoint was not found')
+  expect(text).toContain('Ollama')
+})
+
 test('maps tool_call_incompatible category markers to model/tool guidance', () => {
   const error = APIError.generate(
     400,

src/services/api/errors.ts

@@ -51,7 +51,9 @@ import {
 import { shouldProcessRateLimits } from '../rateLimitMocking.js' // Used for /mock-limits command
 import { extractConnectionErrorDetails, formatAPIError } from './errorUtils.js'
 import {
+  extractOpenAICategoryHost,
   extractOpenAICategoryMarker,
+  isLocalhostLikeHost,
   type OpenAICompatibilityFailureCategory,
 } from './openaiErrorClassification.js'
 
@@ -68,25 +70,29 @@ function mapOpenAICompatibilityFailureToAssistantMessage(options: {
   category: OpenAICompatibilityFailureCategory
   model: string
   rawMessage: string
+  host?: string
 }): AssistantMessage {
   const switchCmd = getIsNonInteractiveSession() ? '--model' : '/model'
   const compactHint = getIsNonInteractiveSession()
     ? 'Reduce prompt size or start a new session.'
     : 'Run /compact or start a new session with /new.'
+  const isLocalhost = options.host === undefined || isLocalhostLikeHost(options.host)
 
   switch (options.category) {
     case 'localhost_resolution_failed':
     case 'connection_refused':
       return createAssistantAPIErrorMessage({
-        content:
-          'Could not connect to the local OpenAI-compatible provider. Ensure the local server is running, then use OPENAI_BASE_URL=http://127.0.0.1:11434/v1 for Ollama.',
+        content: isLocalhost
+          ? 'Could not connect to the local OpenAI-compatible provider. Ensure the local server is running, then use OPENAI_BASE_URL=http://127.0.0.1:11434/v1 for Ollama.'
+          : `Could not connect to the provider at ${options.host}. Verify OPENAI_BASE_URL is correct and that the host is reachable.`,
         error: 'unknown',
       })
 
     case 'endpoint_not_found':
       return createAssistantAPIErrorMessage({
-        content:
-          'Provider endpoint was not found. Confirm OPENAI_BASE_URL targets an OpenAI-compatible /v1 endpoint (for Ollama: http://127.0.0.1:11434/v1).',
+        content: isLocalhost
+          ? 'Provider endpoint was not found. Confirm OPENAI_BASE_URL targets an OpenAI-compatible /v1 endpoint (for Ollama: http://127.0.0.1:11434/v1).'
+          : `Provider endpoint at ${options.host} returned 404. Verify OPENAI_BASE_URL is correct and that the selected model (${options.model}) is supported by this provider.`,
         error: 'invalid_request',
       })
 
@@ -567,6 +573,7 @@ export function getAssistantMessageFromError(
         category: openaiCategory,
         model,
         rawMessage: error.message,
+        host: extractOpenAICategoryHost(error.message),
       })
     }
   }

src/services/api/openaiErrorClassification.test.ts

@@ -4,8 +4,10 @@ import {
   buildOpenAICompatibilityErrorMessage,
   classifyOpenAIHttpFailure,
   classifyOpenAINetworkFailure,
+  extractOpenAICategoryHost,
   extractOpenAICategoryMarker,
   formatOpenAICategoryMarker,
+  isLocalhostLikeHost,
 } from './openaiErrorClassification.js'
 
 test('classifies localhost ECONNREFUSED as connection_refused', () => {
@@ -95,3 +97,58 @@ test('ignores unknown category markers during extraction', () => {
   const malformed = 'OpenAI API error 500 [openai_category=totally_fake_category]'
   expect(extractOpenAICategoryMarker(malformed)).toBeUndefined()
 })
+
+test('endpoint_not_found 404 from a remote host gets a host-aware hint (issue #926)', () => {
+  const failure = classifyOpenAIHttpFailure({
+    status: 404,
+    body: 'Not Found',
+    url: 'https://integrate.api.nvidia.com/v1/chat/completions',
+  })
+
+  expect(failure.category).toBe('endpoint_not_found')
+  expect(failure.requestUrl).toBe('https://integrate.api.nvidia.com/v1/chat/completions')
+  expect(failure.hint).toContain('integrate.api.nvidia.com')
+  expect(failure.hint).not.toContain('local providers')
+})
+
+test('endpoint_not_found 404 from localhost keeps the Ollama-flavored hint', () => {
+  const failure = classifyOpenAIHttpFailure({
+    status: 404,
+    body: 'Not Found',
+    url: 'http://127.0.0.1:11434/v1/chat/completions',
+  })
+
+  expect(failure.category).toBe('endpoint_not_found')
+  expect(failure.hint).toContain('local providers')
+})
+
+test('marker round-trip preserves host segment', () => {
+  const formatted = buildOpenAICompatibilityErrorMessage(
+    'OpenAI API error 404: Not Found',
+    {
+      category: 'endpoint_not_found',
+      hint: 'Endpoint at integrate.api.nvidia.com returned 404.',
+      requestUrl: 'https://integrate.api.nvidia.com/v1/chat/completions',
+    },
+  )
+
+  expect(formatted).toContain('[openai_category=endpoint_not_found,host=integrate.api.nvidia.com]')
+  expect(extractOpenAICategoryMarker(formatted)).toBe('endpoint_not_found')
+  expect(extractOpenAICategoryHost(formatted)).toBe('integrate.api.nvidia.com')
+})
+
+test('marker without host stays backward-compatible', () => {
+  const marker = formatOpenAICategoryMarker('endpoint_not_found')
+  expect(marker).toBe('[openai_category=endpoint_not_found]')
+  expect(extractOpenAICategoryMarker(marker)).toBe('endpoint_not_found')
+  expect(extractOpenAICategoryHost(marker)).toBeUndefined()
+})
+
+test('isLocalhostLikeHost matches loopback variants', () => {
+  expect(isLocalhostLikeHost('localhost')).toBe(true)
+  expect(isLocalhostLikeHost('127.0.0.1')).toBe(true)
+  expect(isLocalhostLikeHost('127.0.0.5')).toBe(true)
+  expect(isLocalhostLikeHost('::1')).toBe(true)
+  expect(isLocalhostLikeHost('integrate.api.nvidia.com')).toBe(false)
+  expect(isLocalhostLikeHost(undefined)).toBe(false)
+})

src/services/api/openaiErrorClassification.ts

@@ -21,6 +21,7 @@ export type OpenAICompatibilityFailure = {
   hint?: string
   code?: string
   status?: number
+  requestUrl?: string
 }
 
 const OPENAI_CATEGORY_MARKER_PREFIX = '[openai_category='
@@ -96,6 +97,11 @@ function isLocalhostLikeHostname(hostname: string | null): boolean {
   return /^127\./.test(hostname)
 }
 
+export function isLocalhostLikeHost(host: string | null | undefined): boolean {
+  if (!host) return false
+  return isLocalhostLikeHostname(host.toLowerCase())
+}
+
 function isContextOverflowMessage(body: string): boolean {
   const lower = body.toLowerCase()
   return (
@@ -149,14 +155,18 @@ function isModelNotFoundMessage(body: string): boolean {
 
 export function formatOpenAICategoryMarker(
   category: OpenAICompatibilityFailureCategory,
+  host?: string,
 ): string {
+  if (host && /^[A-Za-z0-9.\-:]+$/.test(host)) {
+    return `${OPENAI_CATEGORY_MARKER_PREFIX}${category},host=${host}]`
+  }
   return `${OPENAI_CATEGORY_MARKER_PREFIX}${category}]`
 }
 
 export function extractOpenAICategoryMarker(
   message: string,
 ): OpenAICompatibilityFailureCategory | undefined {
-  const match = message.match(/\[openai_category=([a-z_]+)]/)
+  const match = message.match(/\[openai_category=([a-z_]+)(?:,host=[^\]]+)?]/)
   const category = match?.[1]
 
   if (!category || !isOpenAICompatibilityFailureCategory(category)) {
@@ -166,11 +176,17 @@ export function extractOpenAICategoryMarker(
   return category
 }
 
+export function extractOpenAICategoryHost(message: string): string | undefined {
+  const match = message.match(/\[openai_category=[a-z_]+,host=([A-Za-z0-9.\-:]+)]/)
+  return match?.[1]
+}
+
 export function buildOpenAICompatibilityErrorMessage(
   baseMessage: string,
-  failure: Pick<OpenAICompatibilityFailure, 'category' | 'hint'>,
+  failure: Pick<OpenAICompatibilityFailure, 'category' | 'hint' | 'requestUrl'>,
 ): string {
-  const marker = formatOpenAICategoryMarker(failure.category)
+  const host = failure.requestUrl ? getHostname(failure.requestUrl) ?? undefined : undefined
+  const marker = formatOpenAICategoryMarker(failure.category, host)
   const hint = failure.hint ? ` Hint: ${failure.hint}` : ''
   return `${baseMessage} ${marker}${hint}`
 }
@@ -247,8 +263,11 @@ export function classifyOpenAINetworkFailure(
 export function classifyOpenAIHttpFailure(options: {
   status: number
   body: string
+  url?: string
 }): OpenAICompatibilityFailure {
   const body = options.body ?? ''
+  const hostname = options.url ? getHostname(options.url) : null
+  const isLocalHost = isLocalhostLikeHostname(hostname)
 
   if (options.status === 401 || options.status === 403) {
     return {
@@ -284,13 +303,17 @@ export function classifyOpenAIHttpFailure(options: {
   }
 
   if (options.status === 404) {
+    const isRemote = hostname !== null && !isLocalHost
     return {
       source: 'http',
       category: 'endpoint_not_found',
       retryable: false,
       status: options.status,
       message: body,
-      hint: 'Endpoint was not found. Confirm OPENAI_BASE_URL includes /v1 for OpenAI-compatible local providers.',
+      requestUrl: options.url,
+      hint: isRemote
+        ? `Endpoint at ${hostname} returned 404. Verify OPENAI_BASE_URL is correct and the requested model is supported by this provider.`
+        : 'Endpoint was not found. Confirm OPENAI_BASE_URL includes /v1 for OpenAI-compatible local providers.',
     }
   }
 

src/services/api/openaiShim.ts

@@ -1935,7 +1935,9 @@ class OpenAIShimMessages {
         classifyOpenAIHttpFailure({
           status,
           body: errorBody,
+          url: requestUrl,
         })
+      const failureWithUrl = { ...failure, requestUrl: failure.requestUrl ?? requestUrl }
       const redactedUrl = redactUrlForDiagnostics(requestUrl)
 
       logForDebugging(
@@ -1948,7 +1950,7 @@ class OpenAIShimMessages {
         parsedBody,
         buildOpenAICompatibilityErrorMessage(
           `OpenAI API error ${status}: ${errorBody}${rateHint}`,
-          failure,
+          failureWithUrl,
         ),
         responseHeaders,
       )

src/utils/ripgrep.test.ts

@@ -5,15 +5,16 @@ import { resolveRipgrepConfig, wrapRipgrepUnavailableError } from './ripgrep.js'
 
 const MOCK_BUILTIN_PATH = path.normalize(
   process.platform === 'win32'
-    ? `node_modules/@vscode/ripgrep/bin/rg.exe`
-    : `node_modules/@vscode/ripgrep/bin/rg`,
+    ? `vendor/ripgrep/${process.arch}-win32/rg.exe`
+    : `vendor/ripgrep/${process.arch}-${process.platform}/rg`,
 )
 
-test('falls back to system rg when @vscode/ripgrep cannot be resolved', () => {
+test('ripgrepCommand falls back to system rg when builtin binary is missing', () => {
   const config = resolveRipgrepConfig({
     userWantsSystemRipgrep: false,
     bundledMode: false,
-    builtinCommand: null,
+    builtinCommand: MOCK_BUILTIN_PATH,
+    builtinExists: false,
     systemExecutablePath: '/usr/bin/rg',
     processExecPath: '/fake/bun',
   })
@@ -25,11 +26,12 @@ test('falls back to system rg when @vscode/ripgrep cannot be resolved', () => {
   })
 })
 
-test('uses builtin @vscode/ripgrep path when the package resolves', () => {
+test('ripgrepCommand keeps builtin mode when bundled binary exists', () => {
   const config = resolveRipgrepConfig({
     userWantsSystemRipgrep: false,
     bundledMode: false,
     builtinCommand: MOCK_BUILTIN_PATH,
+    builtinExists: true,
     systemExecutablePath: '/usr/bin/rg',
     processExecPath: '/fake/bun',
   })
@@ -41,59 +43,10 @@ test('uses builtin @vscode/ripgrep path when the package resolves', () => {
   })
 })
 
-test('honors USE_BUILTIN_RIPGREP=0 by selecting system rg even when builtin is available', () => {
-  const config = resolveRipgrepConfig({
-    userWantsSystemRipgrep: true,
-    bundledMode: false,
-    builtinCommand: MOCK_BUILTIN_PATH,
-    systemExecutablePath: '/usr/bin/rg',
-    processExecPath: '/fake/bun',
-  })
-
-  expect(config).toMatchObject({
-    mode: 'system',
-    command: 'rg',
-    args: [],
-  })
-})
-
-test('keeps embedded mode for Bun-compiled standalone executables', () => {
-  const config = resolveRipgrepConfig({
-    userWantsSystemRipgrep: false,
-    bundledMode: true,
-    builtinCommand: null,
-    systemExecutablePath: '/usr/bin/rg',
-    processExecPath: '/opt/openclaude/bin/openclaude',
-  })
-
-  expect(config).toMatchObject({
-    mode: 'embedded',
-    command: '/opt/openclaude/bin/openclaude',
-    args: ['--no-config'],
-    argv0: 'rg',
-  })
-})
-
-test('falls through to system rg as a last resort even when not on PATH', () => {
-  const config = resolveRipgrepConfig({
-    userWantsSystemRipgrep: false,
-    bundledMode: false,
-    builtinCommand: null,
-    systemExecutablePath: 'rg',
-    processExecPath: '/fake/bun',
-  })
-
-  expect(config).toMatchObject({
-    mode: 'system',
-    command: 'rg',
-    args: [],
-  })
-})
-
 test('wrapRipgrepUnavailableError explains missing packaged fallback', () => {
   const error = wrapRipgrepUnavailableError(
     { code: 'ENOENT', message: 'spawn rg ENOENT' },
-    { mode: 'builtin', command: 'C:\\fake\\node_modules\\@vscode\\ripgrep\\bin\\rg.exe', args: [] },
+    { mode: 'builtin', command: 'C:\\fake\\vendor\\ripgrep\\rg.exe', args: [] },
     'win32',
   )
 

src/utils/ripgrep.ts

@@ -5,6 +5,7 @@ import memoize from 'lodash-es/memoize.js'
 import { homedir } from 'os'
 import * as path from 'path'
 import { logEvent } from 'src/services/analytics/index.js'
+import { fileURLToPath } from 'url'
 import { isInBundledMode } from './bundledMode.js'
 import { logForDebugging } from './debug.js'
 import { isEnvDefinedFalsy } from './envUtils.js'
@@ -14,6 +15,13 @@ import { logError } from './log.js'
 import { getPlatform } from './platform.js'
 import { countCharInString } from './stringUtils.js'
 
+const __filename = fileURLToPath(import.meta.url)
+// we use node:path.join instead of node:url.resolve because the former doesn't encode spaces
+const __dirname = path.join(
+  __filename,
+  process.env.NODE_ENV === 'test' ? '../../../' : '../',
+)
+
 type RipgrepConfig = {
   mode: 'system' | 'builtin' | 'embedded'
   command: string
@@ -27,31 +35,11 @@ function isErrnoException(error: unknown): error is NodeJS.ErrnoException {
   return error instanceof Error
 }
 
-/**
- * Returns the ripgrep binary path provided by the @vscode/ripgrep package.
- * The package downloads a platform/arch-specific binary at npm install time
- * (cached under the package's bin/ directory). Returns null when the package
- * cannot be resolved — for example when running as a Bun-compiled standalone
- * executable that doesn't ship node_modules.
- */
-function resolveBuiltinRgPath(): string | null {
-  try {
-    // Lazy require so the resolution failure path stays graceful at import
-    // time. The package only exports `rgPath`, so we do not need the rest.
-    const mod = require('@vscode/ripgrep') as { rgPath?: string }
-    if (mod.rgPath && existsSync(mod.rgPath)) {
-      return mod.rgPath
-    }
-  } catch {
-    // Falls through to null — caller decides the fallback.
-  }
-  return null
-}
-
 type ResolveRipgrepConfigArgs = {
   userWantsSystemRipgrep: boolean
   bundledMode: boolean
-  builtinCommand: string | null
+  builtinCommand: string
+  builtinExists: boolean
   systemExecutablePath: string
   processExecPath?: string
 }
@@ -60,6 +48,7 @@ export function resolveRipgrepConfig({
   userWantsSystemRipgrep,
   bundledMode,
   builtinCommand,
+  builtinExists,
   systemExecutablePath,
   processExecPath = process.execPath,
 }: ResolveRipgrepConfigArgs): RipgrepConfig {
@@ -77,7 +66,7 @@ export function resolveRipgrepConfig({
     }
   }
 
-  if (builtinCommand) {
+  if (builtinExists) {
     return { mode: 'builtin', command: builtinCommand, args: [] }
   }
 
@@ -85,9 +74,7 @@ export function resolveRipgrepConfig({
     return { mode: 'system', command: 'rg', args: [] }
   }
 
-  // Last resort — leaves error reporting to the executor when no binary
-  // can be located. wrapRipgrepUnavailableError() surfaces an install hint.
-  return { mode: 'system', command: 'rg', args: [] }
+  return { mode: 'builtin', command: builtinCommand, args: [] }
 }
 
 const getRipgrepConfig = memoize((): RipgrepConfig => {
@@ -95,13 +82,19 @@ const getRipgrepConfig = memoize((): RipgrepConfig => {
     process.env.USE_BUILTIN_RIPGREP,
   )
   const bundledMode = isInBundledMode()
-  const builtinCommand = resolveBuiltinRgPath()
+  const rgRoot = path.resolve(__dirname, 'vendor', 'ripgrep')
+  const builtinCommand =
+    process.platform === 'win32'
+      ? path.resolve(rgRoot, `${process.arch}-win32`, 'rg.exe')
+      : path.resolve(rgRoot, `${process.arch}-${process.platform}`, 'rg')
+  const builtinExists = existsSync(builtinCommand)
   const { cmd: systemExecutablePath } = findExecutable('rg', [])
 
   return resolveRipgrepConfig({
     userWantsSystemRipgrep,
     bundledMode,
     builtinCommand,
+    builtinExists,
     systemExecutablePath,
   })
 })