feat: consolidate docs and blogs into web monorepo#2
Open
hallelx2 wants to merge 21 commits into
Open
Conversation
Brand-matched Fumadocs (Next.js App Router) documentation site for docs.vectorless.store. - Dark theme primary; Geist Sans + Geist Mono + Instrument Serif display - Blue->pink brand gradient and subtle grid background texture - Branded landing page (hero, feature grid, quickstart strip) - Sections: Getting Started, Core Concepts (tree retrieval, no-chunking, treewalk, citations), API Reference (placeholder), SDKs (TS/Python/Go), Self-hosting (Neon + R2 + QStash + Docker, placeholder) - Orama search, OG images, llms.txt - Uses canonical 'treewalk' strategy name (retired 'pageindex')
The Fumadocs site was wrongly built dark-first with a full-page grid wash and aura/glow pseudo-elements. Restyle it to match the brand visual identity in vectorless-why/DESIGN.md, which is LIGHT. - Default theme -> light (RootProvider defaultTheme: 'light'). - Light is now the primary, clean canvas: #FCFCFD, ink #0A0A0A, secondary #3F3F46, muted #71717A, hairlines #E5E7EB. - Remove the heavy body::before grid wash and body::after aura glow. Content/docs pages are clean near-white with whitespace. - Reintroduce only a very faint brand grid rgba(20,86,240,0.06) on the landing hero, masked to fade out (.vl-hero-grid). - Keep the blue->pink signal gradient on exactly one hero line; drop the second serif/gradient register on the pull quote (now plain light sans with a mono eyebrow caption). - Mono eyebrows at 0.16em tracking; body weight ~300 per DESIGN.md. - Quickstart block is now a real dark inset terminal (#0A0A0A) — the one place dark is allowed. - Dark theme retained as an optional toggle, not the default. Restyle only: nav, sidebar, search, section structure and content unchanged. pnpm build passes.
The Fumadocs site was wrongly built dark-first with a full-page grid wash and aura/glow pseudo-elements. Restyle it to match the brand visual identity in vectorless-why/DESIGN.md, which is LIGHT. - Default theme -> light (RootProvider defaultTheme: 'light'). - Light is now the primary, clean canvas: #FCFCFD, ink #0A0A0A, secondary #3F3F46, muted #71717A, hairlines #E5E7EB. - Remove the heavy body::before grid wash and body::after aura glow. Content/docs pages are clean near-white with whitespace. - Reintroduce only a very faint brand grid rgba(20,86,240,0.06) on the landing hero, masked to fade out (.vl-hero-grid). - Keep the blue->pink signal gradient on exactly one hero line; drop the second serif/gradient register on the pull quote (now plain light sans with a mono eyebrow caption). - Mono eyebrows at 0.16em tracking; body weight ~300 per DESIGN.md. - Quickstart block is now a real dark inset terminal (#0A0A0A) — the one place dark is allowed. - Dark theme retained as an optional toggle, not the default. Restyle only: nav, sidebar, search, section structure and content unchanged. pnpm build passes.
Replace the invented node-graph mark with the canonical logo — blue tile + white V (M4 4 L12 20 L20 4) + pink focal dot — matching vectorless-dashboard's icon.tsx; add the matching app favicon. Wordmark in Geist medium. Closes HAL-191.
Replace the invented node-graph mark with the canonical logo — blue tile + white V (M4 4 L12 20 L20 4) + pink focal dot — matching vectorless-dashboard's icon.tsx; add the matching app favicon. Wordmark in Geist medium. Closes HAL-191.
Elevate the marketing/home surface within the Vectorless light brand: - Hero: real V mark, staggered precision reveal, refined terminal inset - New document -> tree -> cited answer SVG motif (hairline, blue path, single pink citation dot; edges draw + nodes snap) - Feature grid with mono index numbers and hover accent - Parse/Walk/Cite three-step band - Refined Instrument-Serif thesis closer with CTA No new gradients, no dark page bg, no new mark - honors DESIGN.md.
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Reviewer's GuideAdds new apps/docs and apps/blogs Next.js applications into the web monorepo (Fumadocs-based docs site and a rich marketing/blog experience), wires them into shared branding/LLM tooling, and removes deprecated SDK/openapi/mcp-tools packages from the workspace. File-Level Changes
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
📝 WalkthroughWalkthroughThis PR adds two new Next.js apps ( ChangesMonorepo Maintenance: CI, React Pins, and Legacy Package Removal
New apps/blogs AI Demo Application
New apps/docs Fumadocs Documentation Site
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Hey - I've found 1 issue, and left some high level feedback:
- The blog landing page (
apps/blogs/app/page.tsx) has grown very large and mixes data, layout, and interactive demo logic in a single component; consider extracting the blog post data, vector illustrations, and the retrieval demo into separate modules/hooks to keep the page component focused and easier to reason about. - The Vectorless logo/mark is implemented in multiple places (
apps/docs/src/lib/layout.shared.tsx,(home)/page.tsx,apps/web, and blog); it would be more maintainable to centralize this into a shared component so future brand tweaks only need to be made in one place. - The new blog app’s
package.jsonpulls in several dependencies that aren’t used by the current codebase (e.g.@google/genai,@hookform/resolvers,firebase-tools); trimming these will keep installs faster and reduce maintenance surface area.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- The blog landing page (`apps/blogs/app/page.tsx`) has grown very large and mixes data, layout, and interactive demo logic in a single component; consider extracting the blog post data, vector illustrations, and the retrieval demo into separate modules/hooks to keep the page component focused and easier to reason about.
- The Vectorless logo/mark is implemented in multiple places (`apps/docs/src/lib/layout.shared.tsx`, `(home)/page.tsx`, `apps/web`, and blog); it would be more maintainable to centralize this into a shared component so future brand tweaks only need to be made in one place.
- The new blog app’s `package.json` pulls in several dependencies that aren’t used by the current codebase (e.g. `@google/genai`, `@hookform/resolvers`, `firebase-tools`); trimming these will keep installs faster and reduce maintenance surface area.
## Individual Comments
### Comment 1
<location path="apps/blogs/next.config.ts" line_range="25-26" />
<code_context>
+ output: process.env.NEXT_STANDALONE === 'true' ? 'standalone' : undefined,
+ transpilePackages: ['motion'],
+ webpack: (config, {dev}) => {
+ // HMR is disabled in AI Studio via DISABLE_HMR env var.
+ // Do not modify�file watching is disabled to prevent flickering during agent edits.
+ if (dev && process.env.DISABLE_HMR === 'true') {
+ config.watchOptions = {
</code_context>
<issue_to_address>
**nitpick (typo):** Fix the mojibake em dash in the webpack comment to avoid encoding noise.
The line `// Do not modifyâ��file watching is disabled...` contains a mis-encoded em dash. Please change it to `// Do not modify—file watching is disabled...` to keep the comment readable and avoid encoding artifacts.
```suggestion
// HMR is disabled in AI Studio via DISABLE_HMR env var.
// Do not modify—file watching is disabled to prevent flickering during agent edits.
```
</issue_to_address>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
Comment on lines
+25
to
+26
| // HMR is disabled in AI Studio via DISABLE_HMR env var. | ||
| // Do not modifyâfile watching is disabled to prevent flickering during agent edits. |
There was a problem hiding this comment.
nitpick (typo): Fix the mojibake em dash in the webpack comment to avoid encoding noise.
The line // Do not modifyâ��file watching is disabled... contains a mis-encoded em dash. Please change it to // Do not modify—file watching is disabled... to keep the comment readable and avoid encoding artifacts.
Suggested change
| // HMR is disabled in AI Studio via DISABLE_HMR env var. | |
| // Do not modify�file watching is disabled to prevent flickering during agent edits. | |
| // HMR is disabled in AI Studio via DISABLE_HMR env var. | |
| // Do not modify—file watching is disabled to prevent flickering during agent edits. |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 15
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
.github/workflows/ci.yml (1)
10-22:⚠️ Potential issue | 🟠 Major | ⚡ Quick winAdd explicit permissions block to follow least-privilege principle.
The workflow uses default permissions which are overly broad. GitHub Actions should explicitly declare minimal required permissions.
🔒 Proposed fix to add permissions block
lint-and-typecheck: name: Lint & Typecheck runs-on: ubuntu-latest + permissions: + contents: read steps:🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.github/workflows/ci.yml around lines 10 - 22, The lint-and-typecheck workflow job is missing an explicit permissions block, which allows overly broad default permissions. Add a permissions block to the lint-and-typecheck job that specifies only the minimal required permissions. For this job which only needs to checkout and read the repository contents, add a permissions block with contents set to read. This follows the least-privilege principle by explicitly declaring only the necessary permissions required for the job to execute.Source: Linters/SAST tools
🧹 Nitpick comments (6)
apps/blogs/app/page.tsx (2)
614-618: 💤 Low valueAdd
referrerPolicy="no-referrer"for consistency.The author avatar image at line 520-523 includes
referrerPolicy="no-referrer", but the same image element in the article drawer (line 615) is missing it. Apply the same policy for consistency.🔧 Proposed fix
<img src={activeArticle.author.avatarUrl} alt={activeArticle.author.name} className="object-cover w-full h-full" + referrerPolicy="no-referrer" />🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@apps/blogs/app/page.tsx` around lines 614 - 618, The img element displaying activeArticle.author.avatarUrl in the article drawer section is missing the referrerPolicy attribute for consistency. Add the referrerPolicy="no-referrer" attribute to the img element with className "object-cover w-full h-full" to match the implementation of the same avatar image element that appears earlier in the component at line 520-523.
699-699: ⚡ Quick winAvoid
as anytype assertion.The
as anycast bypasses TypeScript's type safety. Instead, define a union type or use a type-safe assertion.🔧 Proposed fix
+type DocId = 'vectorless-whitepaper.pdf' | 'fastify-manual.md' | 'terms-of-service.docx'; + const [selectedDoc, setSelectedDoc] = useState<'vectorless-whitepaper.pdf' | 'fastify-manual.md' | 'terms-of-service.docx'>('vectorless-whitepaper.pdf'); // ... - onClick={() => setSelectedDoc(doc.id as any)} + onClick={() => setSelectedDoc(doc.id as DocId)}Or refactor the document list to be typed:
{[ - { id: 'vectorless-whitepaper.pdf', size: '2.4 MB', type: 'PDF' }, - { id: 'fastify-manual.md', size: '480 KB', type: 'Markdown' }, - { id: 'terms-of-service.docx', size: '1.1 MB', type: 'Word Doc' } + { id: 'vectorless-whitepaper.pdf' as const, size: '2.4 MB', type: 'PDF' }, + { id: 'fastify-manual.md' as const, size: '480 KB', type: 'Markdown' }, + { id: 'terms-of-service.docx' as const, size: '1.1 MB', type: 'Word Doc' } ].map((doc) => {🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@apps/blogs/app/page.tsx` at line 699, Remove the `as any` type assertion from the onClick handler where setSelectedDoc is called with doc.id. Instead, examine the actual type of doc.id and the expected parameter type of setSelectedDoc function. Either refactor the document list to ensure doc.id has the correct type that matches setSelectedDoc's parameter type, or define a union type that safely captures the possible types of doc.id without resorting to the unsafe `as any` cast.apps/blogs/package.json (1)
15-15: 💤 Low valueConsider removing
autoprefixer(redundant with Tailwind v4).Tailwind CSS v4 includes built-in Lightning CSS integration that handles vendor prefixes automatically. The separate
autoprefixerdependency is likely redundant unless you have non-Tailwind CSS that requires it.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@apps/blogs/package.json` at line 15, Remove the autoprefixer dependency from the package.json file by deleting the line containing "autoprefixer": "^10.4.21" in the dependencies section. Verify that your project is using Tailwind CSS v4 (check tailwind.config.js or similar configuration file to confirm it's v4) and that there are no custom CSS files outside of Tailwind that require autoprefixer functionality. After removal, you may need to run npm install or your package manager's equivalent to update the lock file and ensure Tailwind's built-in vendor prefixing handles all necessary browser compatibility.apps/blogs/hooks/use-mobile.ts (1)
5-21: ⚖️ Poor tradeoffPotential SSR hydration mismatch and inconsistent breakpoint logic.
The hook has several issues:
Hydration mismatch risk: Initial state is
undefined, coerced tofalsevia!!isMobile(line 20). During SSR or initial render, this returnsfalse, but aftersetTimeoutfires, it may switch totrue, causing a visual flicker.Inconsistent breakpoint checks: The
matchMedialistener useswindow.innerWidth < MOBILE_BREAKPOINT(line 11), but should usemql.matchesto stay consistent with the media query definition.Missing SSR guard: If this hook is called during SSR, accessing
window(line 9) will throw a runtime error.Unnecessary
setTimeout: The zero-delay timeout is a workaround that indicates a deeper architectural issue.♻️ Recommended fix
export function useIsMobile() { - const [isMobile, setIsMobile] = React.useState<boolean | undefined>(undefined) + const [isMobile, setIsMobile] = React.useState<boolean | undefined>( + typeof window !== 'undefined' + ? window.innerWidth < MOBILE_BREAKPOINT + : undefined + ) React.useEffect(() => { const mql = window.matchMedia(`(max-width: ${MOBILE_BREAKPOINT - 1}px)`) const onChange = () => { - setIsMobile(window.innerWidth < MOBILE_BREAKPOINT) + setIsMobile(mql.matches) } mql.addEventListener("change", onChange) - setTimeout(() => { - setIsMobile(window.innerWidth < MOBILE_BREAKPOINT) - }, 0) + setIsMobile(mql.matches) return () => mql.removeEventListener("change", onChange) }, []) return !!isMobile }Alternatively, consider using a library like
react-responsivefor more robust media query hooks.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@apps/blogs/hooks/use-mobile.ts` around lines 5 - 21, The useIsMobile hook has SSR safety and hydration consistency issues. First, add a guard to check if window exists before accessing it to prevent SSR errors. Second, replace the window.innerWidth comparison with mql.matches to ensure consistent breakpoint logic between the matchMedia listener and initial state. Third, initialize isMobile to false instead of undefined to prevent hydration mismatches between server and client renders. Fourth, remove the setTimeout call as it's unnecessary when using mql.matches for state synchronization, and ensure the return statement properly coerces the boolean value.apps/blogs/postcss.config.mjs (1)
3-6: ⚡ Quick winDrop
autoprefixerunless you still target legacy browsers.Tailwind 4's PostCSS pipeline already handles prefixing, so this extra plugin is redundant for the modern app stack here.
Suggested cleanup
plugins: { '`@tailwindcss/postcss`': {}, - autoprefixer: {}, },🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@apps/blogs/postcss.config.mjs` around lines 3 - 6, Remove the autoprefixer plugin from the plugins object in the postcss.config.mjs file. Since Tailwind 4's PostCSS pipeline already handles CSS prefixing automatically, the redundant autoprefixer entry should be deleted, leaving only the '`@tailwindcss/postcss`' plugin in the configuration.apps/docs/src/lib/layout.shared.tsx (1)
51-51: 💤 Low valueGitHub URL inconsistency with gitConfig.
The
githubUrlhardcodesvectorless-engineas the repository name, butgitConfig.repo(fromshared.ts) is set tovectorless-docs. This creates an inconsistency wheregitConfigis defined but not used consistently.If the intent is to link to the engine repository from the docs site, consider either:
- Renaming
gitConfig.repotogitConfig.docsRepoand addinggitConfig.engineRepo, or- Adding a comment explaining why
gitConfig.repodiffers from the UI links♻️ Proposed refactor for clarity
export const gitConfig = { user: 'hallelx2', - repo: 'vectorless-docs', + docsRepo: 'vectorless-docs', + engineRepo: 'vectorless-engine', branch: 'main', };Then in layout.shared.tsx:
- githubUrl: `https://github.com/${gitConfig.user}/vectorless-engine`, + githubUrl: `https://github.com/${gitConfig.user}/${gitConfig.engineRepo}`,And update the Engine link:
- url: 'https://github.com/hallelx2/vectorless-engine', + url: `https://github.com/${gitConfig.user}/${gitConfig.engineRepo}`,🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@apps/docs/src/lib/layout.shared.tsx` at line 51, The githubUrl hardcodes 'vectorless-engine' as the repository name, but this is inconsistent with gitConfig.repo which is set to 'vectorless-docs' in shared.ts. To fix this inconsistency, refactor the gitConfig object in shared.ts to distinguish between the docs repository and engine repository by renaming gitConfig.repo to gitConfig.docsRepo and adding a new gitConfig.engineRepo property set to 'vectorless-engine'. Then update the githubUrl assignment in layout.shared.tsx to use gitConfig.engineRepo instead of the hardcoded 'vectorless-engine' string.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@apps/blogs/.env.example`:
- Line 4: Remove the double quotes from the example secret values in the
.env.example file. Update GEMINI_API_KEY and the other API key placeholder
mentioned (line 9) by removing the surrounding quotes so the raw values are
shown without quotes. This prevents users from accidentally copying the quotes
as part of their actual secrets when pasting values into their .env
configuration.
In `@apps/blogs/metadata.json`:
- Line 5: The metadata.json file falsely claims the
MAJOR_CAPABILITY_SERVER_SIDE_GEMINI_API capability, but the actual
implementation in the blog app uses mocked responses in the handleRetrieve
function rather than any real Gemini API integration. Remove the
MAJOR_CAPABILITY_SERVER_SIDE_GEMINI_API entry from the majorCapabilities array
in metadata.json to accurately reflect that the app does not use server-side
Gemini API functionality.
In `@apps/blogs/next.config.ts`:
- Line 22: The conditional output logic in next.config.ts expects the
NEXT_STANDALONE environment variable to be set to 'true' to generate
.next/standalone, but this variable is not being configured in the CI and
deployment environments. Add NEXT_STANDALONE=true to the build environment in
.github/workflows/ci.yml and to the relevant service configurations in
docker-compose files to ensure the standalone output mode is actually generated
during builds and available for deployments that depend on it.
In `@apps/blogs/README.md`:
- Around line 1-20: Replace the AI Studio template boilerplate content in the
README with documentation specific to the Vectorless blog app. Remove the Google
AI Studio app URL reference (line 9), update the environment setup instructions
to reference .env.example instead of .env.local (line 18) and clarify that the
app does not require a Gemini API key, and replace the AI Studio banner image
(line 2) with appropriate Vectorless branding to match the actual application
purpose and technology stack.
In `@apps/docs/content/docs/index.mdx`:
- Around line 48-50: The description property in the API Reference card
(title="API Reference" with href="/docs/api") currently states the API reference
is "generated from the engine's OpenAPI spec", but since this PR removes the
packages/openapi package, this description is now outdated and misleading.
Update the description text to accurately reflect the current source and purpose
of the API reference without referencing the OpenAPI spec generation.
In `@apps/docs/content/docs/sdks/go.mdx`:
- Around line 13-15: The Go module import path
`github.com/hallelx2/vectorless-go` in the go get command is invalid and returns
a 404 error from the module proxy, causing the installation to fail for users.
Verify the correct and currently published module path for the Go SDK and update
the go get command with the valid module path in the code snippet.
In `@apps/docs/content/docs/sdks/typescript.mdx`:
- Around line 45-57: In the streaming example under the Streaming section,
replace the Node.js-specific process.stdout.write call with a runtime-agnostic
approach such as console.log to handle the token text output. This ensures the
example works consistently across both Node.js and edge runtime environments
without requiring separate examples for different runtimes.
In `@apps/docs/content/docs/self-hosting/index.mdx`:
- Line 55: The docker run command uses the mutable `:latest` tag for the
ghcr.io/hallelx2/vectorless-engine image reference, which prevents reproducible
deployments and makes rollbacks difficult. Replace the `:latest` tag with a
specific version number (e.g., `:v1.0.0` or `:0.1.0`) or a digest hash to pin
the exact image being deployed. This ensures consistent, reproducible behavior
across deployments and enables safe version rollbacks when needed.
In `@apps/docs/proxy.ts`:
- Around line 14-29: The proxy function in proxy.ts is unused and needs to be
integrated into Next.js's middleware pipeline. Rename proxy.ts to middleware.ts
and move it to either the project root directory or the src/ directory
(depending on your project structure) so Next.js 16 can recognize and
automatically load it as middleware. If the proxy function is no longer needed
and won't be used, simply delete the proxy.ts file instead.
In `@apps/docs/README.md`:
- Around line 36-41: In the Design section of the README, update the theme
description that currently states "Dark theme primary" to accurately reflect
that the docs app is light-first with an optional dark toggle. Replace the
inaccurate theme designation with text that correctly describes the actual theme
implementation and user experience of the application.
In `@apps/docs/src/app/`(home)/page.tsx:
- Line 292: The Link component with the text "Status" currently uses href="#" as
a placeholder, which creates a dead navigation link in production. Either
replace the href="#" with a real destination URL pointing to a valid status
page, or if this should not be a navigation link at all, convert it from a Link
component to regular HTML text wrapped in a span or div element while preserving
the existing className styling.
In `@apps/docs/src/app/global.css`:
- Line 6: The Tailwind source scanning configuration in global.css only targets
ts and tsx files, which excludes content/**.mdx documentation files from being
scanned for utility classes, causing those classes to be purged from the output.
Modify the `@source` directive to include the content directory and mdx files by
updating the pattern to also scan documentation markdown files alongside the
existing TypeScript sources.
In `@apps/docs/src/app/layout.tsx`:
- Around line 21-24: The theme configuration object in the layout component has
theming disabled with enabled set to false and forcedTheme set to light, which
prevents users from switching between light and dark themes. To enable theme
switching functionality, change the enabled property to true and adjust the
forcedTheme setting appropriately to allow the theme toggle to work. This will
restore the ability for users to switch themes if dark mode is intended as an
optional choice.
In `@apps/docs/src/app/llms.txt/route.ts`:
- Around line 6-8: The GET function returns a Response without an explicit
Content-Type header, which causes the endpoint to be unreliable for clients
expecting plain-text. Modify the Response constructor in the GET function to
include a headers object with the Content-Type set to "text/plain" so that
clients can properly identify the response format.
In `@apps/docs/src/app/og/docs/`[...slug]/route.tsx:
- Around line 9-27: In the generateStaticParams() function, remove the lang:
page.locale property from the returned object since the route pattern
/og/docs/[...slug] does not have a [lang] segment. The function should only
return slug: getPageImage(page).segments. Additionally, investigate why
slug.slice(0, -1) is needed in the GET function while
getPageImage(page).segments appends 'image.png', and either clarify or fix this
contract mismatch between the helper function and route expectations.
---
Outside diff comments:
In @.github/workflows/ci.yml:
- Around line 10-22: The lint-and-typecheck workflow job is missing an explicit
permissions block, which allows overly broad default permissions. Add a
permissions block to the lint-and-typecheck job that specifies only the minimal
required permissions. For this job which only needs to checkout and read the
repository contents, add a permissions block with contents set to read. This
follows the least-privilege principle by explicitly declaring only the necessary
permissions required for the job to execute.
---
Nitpick comments:
In `@apps/blogs/app/page.tsx`:
- Around line 614-618: The img element displaying activeArticle.author.avatarUrl
in the article drawer section is missing the referrerPolicy attribute for
consistency. Add the referrerPolicy="no-referrer" attribute to the img element
with className "object-cover w-full h-full" to match the implementation of the
same avatar image element that appears earlier in the component at line 520-523.
- Line 699: Remove the `as any` type assertion from the onClick handler where
setSelectedDoc is called with doc.id. Instead, examine the actual type of doc.id
and the expected parameter type of setSelectedDoc function. Either refactor the
document list to ensure doc.id has the correct type that matches
setSelectedDoc's parameter type, or define a union type that safely captures the
possible types of doc.id without resorting to the unsafe `as any` cast.
In `@apps/blogs/hooks/use-mobile.ts`:
- Around line 5-21: The useIsMobile hook has SSR safety and hydration
consistency issues. First, add a guard to check if window exists before
accessing it to prevent SSR errors. Second, replace the window.innerWidth
comparison with mql.matches to ensure consistent breakpoint logic between the
matchMedia listener and initial state. Third, initialize isMobile to false
instead of undefined to prevent hydration mismatches between server and client
renders. Fourth, remove the setTimeout call as it's unnecessary when using
mql.matches for state synchronization, and ensure the return statement properly
coerces the boolean value.
In `@apps/blogs/package.json`:
- Line 15: Remove the autoprefixer dependency from the package.json file by
deleting the line containing "autoprefixer": "^10.4.21" in the dependencies
section. Verify that your project is using Tailwind CSS v4 (check
tailwind.config.js or similar configuration file to confirm it's v4) and that
there are no custom CSS files outside of Tailwind that require autoprefixer
functionality. After removal, you may need to run npm install or your package
manager's equivalent to update the lock file and ensure Tailwind's built-in
vendor prefixing handles all necessary browser compatibility.
In `@apps/blogs/postcss.config.mjs`:
- Around line 3-6: Remove the autoprefixer plugin from the plugins object in the
postcss.config.mjs file. Since Tailwind 4's PostCSS pipeline already handles CSS
prefixing automatically, the redundant autoprefixer entry should be deleted,
leaving only the '`@tailwindcss/postcss`' plugin in the configuration.
In `@apps/docs/src/lib/layout.shared.tsx`:
- Line 51: The githubUrl hardcodes 'vectorless-engine' as the repository name,
but this is inconsistent with gitConfig.repo which is set to 'vectorless-docs'
in shared.ts. To fix this inconsistency, refactor the gitConfig object in
shared.ts to distinguish between the docs repository and engine repository by
renaming gitConfig.repo to gitConfig.docsRepo and adding a new
gitConfig.engineRepo property set to 'vectorless-engine'. Then update the
githubUrl assignment in layout.shared.tsx to use gitConfig.engineRepo instead of
the hardcoded 'vectorless-engine' string.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: da21b349-7a1f-4455-b0e5-75419ceb58b3
⛔ Files ignored due to path filters (9)
apps/blogs/src/assets/images/golf_bag_coat_1781964385599.jpgis excluded by!**/*.jpgapps/blogs/src/assets/images/golf_bg_1781964406657.jpgis excluded by!**/*.jpgapps/blogs/src/assets/images/golf_lying_grass_1781964368724.jpgis excluded by!**/*.jpgapps/docs/src/fonts/Geist.woff2is excluded by!**/*.woff2apps/docs/src/fonts/GeistMono.woff2is excluded by!**/*.woff2apps/docs/src/fonts/InstrumentSerif-Italic.ttfis excluded by!**/*.ttfapps/docs/src/fonts/InstrumentSerif.ttfis excluded by!**/*.ttfapps/web/public/Vectorless-Whitepaper.pdfis excluded by!**/*.pdfpnpm-lock.yamlis excluded by!**/pnpm-lock.yaml
📒 Files selected for processing (96)
.github/workflows/ci.yml.github/workflows/publish-ts-sdk.ymlapps/blogs/.env.exampleapps/blogs/.eslintrc.jsonapps/blogs/.gitignoreapps/blogs/README.mdapps/blogs/app/globals.cssapps/blogs/app/layout.tsxapps/blogs/app/page.tsxapps/blogs/eslint.config.mjsapps/blogs/hooks/use-mobile.tsapps/blogs/lib/utils.tsapps/blogs/metadata.jsonapps/blogs/next-env.d.tsapps/blogs/next.config.tsapps/blogs/package.jsonapps/blogs/postcss.config.mjsapps/blogs/tsconfig.jsonapps/docs/.gitignoreapps/docs/README.mdapps/docs/content/docs/api/index.mdxapps/docs/content/docs/api/meta.jsonapps/docs/content/docs/concepts/citations.mdxapps/docs/content/docs/concepts/index.mdxapps/docs/content/docs/concepts/meta.jsonapps/docs/content/docs/concepts/no-chunking.mdxapps/docs/content/docs/concepts/tree-retrieval.mdxapps/docs/content/docs/concepts/treewalk.mdxapps/docs/content/docs/getting-started/index.mdxapps/docs/content/docs/getting-started/meta.jsonapps/docs/content/docs/getting-started/quickstart.mdxapps/docs/content/docs/index.mdxapps/docs/content/docs/meta.jsonapps/docs/content/docs/sdks/go.mdxapps/docs/content/docs/sdks/index.mdxapps/docs/content/docs/sdks/meta.jsonapps/docs/content/docs/sdks/python.mdxapps/docs/content/docs/sdks/typescript.mdxapps/docs/content/docs/self-hosting/index.mdxapps/docs/content/docs/self-hosting/meta.jsonapps/docs/eslint.config.mjsapps/docs/next.config.mjsapps/docs/package.jsonapps/docs/postcss.config.mjsapps/docs/proxy.tsapps/docs/source.config.tsapps/docs/src/app/(home)/layout.tsxapps/docs/src/app/(home)/page.tsxapps/docs/src/app/(home)/treewalk-motif.tsxapps/docs/src/app/api/search/route.tsapps/docs/src/app/docs/[[...slug]]/page.tsxapps/docs/src/app/docs/layout.tsxapps/docs/src/app/global.cssapps/docs/src/app/icon.tsxapps/docs/src/app/layout.tsxapps/docs/src/app/llms-full.txt/route.tsapps/docs/src/app/llms.mdx/docs/[[...slug]]/route.tsapps/docs/src/app/llms.txt/route.tsapps/docs/src/app/og/docs/[...slug]/route.tsxapps/docs/src/components/HeroShader.tsxapps/docs/src/components/mdx.tsxapps/docs/src/hooks/useGsap.tsapps/docs/src/lib/cn.tsapps/docs/src/lib/fonts.tsapps/docs/src/lib/layout.shared.tsxapps/docs/src/lib/shared.tsapps/docs/src/lib/source.tsapps/docs/tsconfig.jsonapps/web/package.jsonapps/web/tsconfig.tsbuildinfopackage.jsonpackages/mcp-tools/package.jsonpackages/mcp-tools/src/handlers.tspackages/mcp-tools/src/index.tspackages/mcp-tools/src/schemas.tspackages/mcp-tools/src/tools.tspackages/mcp-tools/src/types.tspackages/mcp-tools/tsconfig.jsonpackages/mcp-tools/tsup.config.tspackages/openapi/openapi.yamlpackages/openapi/package.jsonpackages/ts-sdk/LICENSEpackages/ts-sdk/README.mdpackages/ts-sdk/__tests__/client.test.tspackages/ts-sdk/__tests__/integration.test.tspackages/ts-sdk/package.jsonpackages/ts-sdk/src/client.tspackages/ts-sdk/src/config.tspackages/ts-sdk/src/errors.tspackages/ts-sdk/src/http.tspackages/ts-sdk/src/index.tspackages/ts-sdk/src/retry.tspackages/ts-sdk/src/types.tspackages/ts-sdk/src/upload.tspackages/ts-sdk/tsconfig.jsonpackages/ts-sdk/tsup.config.ts
💤 Files with no reviewable changes (26)
- packages/ts-sdk/tsconfig.json
- packages/ts-sdk/tests/client.test.ts
- packages/mcp-tools/package.json
- packages/mcp-tools/src/types.ts
- packages/ts-sdk/tests/integration.test.ts
- packages/ts-sdk/tsup.config.ts
- packages/ts-sdk/src/config.ts
- packages/ts-sdk/src/http.ts
- packages/ts-sdk/package.json
- packages/ts-sdk/src/upload.ts
- packages/mcp-tools/src/tools.ts
- packages/ts-sdk/LICENSE
- packages/mcp-tools/src/handlers.ts
- packages/ts-sdk/README.md
- packages/ts-sdk/src/retry.ts
- packages/mcp-tools/src/schemas.ts
- packages/ts-sdk/src/index.ts
- packages/ts-sdk/src/errors.ts
- packages/mcp-tools/tsup.config.ts
- .github/workflows/publish-ts-sdk.yml
- packages/ts-sdk/src/client.ts
- packages/openapi/openapi.yaml
- packages/openapi/package.json
- packages/ts-sdk/src/types.ts
- packages/mcp-tools/src/index.ts
- packages/mcp-tools/tsconfig.json
| # GEMINI_API_KEY: Required for Gemini AI API calls. | ||
| # AI Studio automatically injects this at runtime from user secrets. | ||
| # Users configure this via the Secrets panel in the AI Studio UI. | ||
| GEMINI_API_KEY="MY_GEMINI_API_KEY" |
There was a problem hiding this comment.
Remove the quotes from the example secrets.
These placeholders are meant to be copied into Secrets / .env as raw values. Keeping the quotes makes them easier to paste incorrectly, and dotenv-linter already flags both lines.
Suggested cleanup
-GEMINI_API_KEY="MY_GEMINI_API_KEY"
+GEMINI_API_KEY=MY_GEMINI_API_KEY
# APP_URL: The URL where this applet is hosted.
@@
-APP_URL="MY_APP_URL"
+APP_URL=MY_APP_URLAlso applies to: 9-9
🧰 Tools
🪛 dotenv-linter (4.0.0)
[warning] 4-4: [QuoteCharacter] The value has quote characters (', ")
(QuoteCharacter)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@apps/blogs/.env.example` at line 4, Remove the double quotes from the example
secret values in the .env.example file. Update GEMINI_API_KEY and the other API
key placeholder mentioned (line 9) by removing the surrounding quotes so the raw
values are shown without quotes. This prevents users from accidentally copying
the quotes as part of their actual secrets when pasting values into their .env
configuration.
Source: Linters/SAST tools
| "name": "Vectorless Blog", | ||
| "description": "Insights, specifications, and performance analyses of structure-preserving retrieval architectures.", | ||
| "requestFramePermissions": [], | ||
| "majorCapabilities": ["MAJOR_CAPABILITY_SERVER_SIDE_GEMINI_API"] |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Search for actual `@google/genai` imports and usage in apps/blogs
rg -n --type=ts --type=tsx -C3 "from ['\"]`@google/genai`['\"]" apps/blogs/
rg -n --type=ts --type=tsx -C3 "GoogleGenerativeAI|GenerativeModel" apps/blogs/Repository: hallelx2/vectorless
Length of output: 119
🏁 Script executed:
# Search for `@google/genai` imports and usage - corrected approach
rg -n "`@google/genai`" apps/blogs/ -C3
rg -n "GoogleGenerativeAI|GenerativeModel" apps/blogs/ -C3Repository: hallelx2/vectorless
Length of output: 445
🏁 Script executed:
# Read metadata.json to confirm the capability claim
cat apps/blogs/metadata.jsonRepository: hallelx2/vectorless
Length of output: 316
🏁 Script executed:
# Read app/page.tsx to check for actual Gemini API usage
wc -l apps/blogs/app/page.tsx && head -50 apps/blogs/app/page.tsxRepository: hallelx2/vectorless
Length of output: 1123
🏁 Script executed:
# Search for GoogleGenerativeAI or GenerativeModel usage
rg -n "GoogleGenerativeAI|GenerativeModel" apps/blogs/app/page.tsx -C5Repository: hallelx2/vectorless
Length of output: 45
🏁 Script executed:
# Search for setTimeout/mock implementations in page.tsx
rg -n "setTimeout|mock" apps/blogs/app/page.tsx -C3Repository: hallelx2/vectorless
Length of output: 1264
🏁 Script executed:
# Check imports at the top of page.tsx
sed -n '1,100p' apps/blogs/app/page.tsx | grep -n "import"Repository: hallelx2/vectorless
Length of output: 195
🏁 Script executed:
# Get complete imports section
sed -n '1,30p' apps/blogs/app/page.tsxRepository: hallelx2/vectorless
Length of output: 599
🏁 Script executed:
# Check the mock retrieval logic more thoroughly
sed -n '280,330p' apps/blogs/app/page.tsxRepository: hallelx2/vectorless
Length of output: 2457
Capability claim inconsistent with implementation.
metadata.json claims MAJOR_CAPABILITY_SERVER_SIDE_GEMINI_API, but the blog app's demo uses entirely mocked retrieval logic. The @google/genai dependency is listed in package.json but is never imported or used. The handleRetrieve function in app/page.tsx (lines 280–330) uses hardcoded responses triggered by setTimeout, with no actual Gemini API integration.
Remove the capability claim from metadata.json or implement actual server-side Gemini API usage.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@apps/blogs/metadata.json` at line 5, The metadata.json file falsely claims
the MAJOR_CAPABILITY_SERVER_SIDE_GEMINI_API capability, but the actual
implementation in the blog app uses mocked responses in the handleRetrieve
function rather than any real Gemini API integration. Remove the
MAJOR_CAPABILITY_SERVER_SIDE_GEMINI_API entry from the majorCapabilities array
in metadata.json to accurately reflect that the app does not use server-side
Gemini API functionality.
| }, | ||
| ], | ||
| }, | ||
| output: process.env.NEXT_STANDALONE === 'true' ? 'standalone' : undefined, |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
rg -n "NEXT_STANDALONE|standalone" .Repository: hallelx2/vectorless
Length of output: 280
🏁 Script executed:
find . -type f \( -name "*.yml" -o -name "*.yaml" -o -name "Dockerfile*" -o -name "docker-compose*" -o -name "*.json" \) | grep -E "(workflows|docker|deploy|build)" | head -20Repository: hallelx2/vectorless
Length of output: 259
🏁 Script executed:
fd -t f -e yml -e yaml . .github/workflows/ 2>/dev/null || echo "No workflows found"Repository: hallelx2/vectorless
Length of output: 125
🏁 Script executed:
fd -t f \( -name "Dockerfile" -o -name "docker-compose*" -o -name "*.yml" -o -name "*.yaml" \) . | head -30Repository: hallelx2/vectorless
Length of output: 294
🏁 Script executed:
cat -n package.json | grep -A 5 -B 5 "build\|deploy" || echo "Check package.json manually"Repository: hallelx2/vectorless
Length of output: 432
🏁 Script executed:
cat .github/workflows/ci.ymlRepository: hallelx2/vectorless
Length of output: 1404
🏁 Script executed:
cat ./docker-compose.ymlRepository: hallelx2/vectorless
Length of output: 463
🏁 Script executed:
cat ./deploy/local/docker-compose.ymlRepository: hallelx2/vectorless
Length of output: 5719
CI and deploy configurations must set NEXT_STANDALONE=true if standalone output is required.
The conditional logic in apps/blogs/next.config.ts expects NEXT_STANDALONE=true to generate .next/standalone, but neither the CI workflow (.github/workflows/ci.yml) nor the docker-compose files set this variable. When the flag is omitted, the build silently falls back to the default output layout. Any deployment expecting .next/standalone will fail without explicit environment variable configuration.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@apps/blogs/next.config.ts` at line 22, The conditional output logic in
next.config.ts expects the NEXT_STANDALONE environment variable to be set to
'true' to generate .next/standalone, but this variable is not being configured
in the CI and deployment environments. Add NEXT_STANDALONE=true to the build
environment in .github/workflows/ci.yml and to the relevant service
configurations in docker-compose files to ensure the standalone output mode is
actually generated during builds and available for deployments that depend on
it.
Comment on lines
+1
to
+20
| <div align="center"> | ||
| <img width="1200" height="475" alt="GHBanner" src="https://ai.google.dev/static/site-assets/images/share-ais-513315318.png" /> | ||
| </div> | ||
|
|
||
| # Run and deploy your AI Studio app | ||
|
|
||
| This contains everything you need to run your app locally. | ||
|
|
||
| View your app in AI Studio: https://ai.studio/apps/3b8c6a97-eb59-4fb0-aade-3976e82644b1 | ||
|
|
||
| ## Run Locally | ||
|
|
||
| **Prerequisites:** Node.js | ||
|
|
||
|
|
||
| 1. Install dependencies: | ||
| `npm install` | ||
| 2. Set the `GEMINI_API_KEY` in [.env.local](.env.local) to your Gemini API key | ||
| 3. Run the app: | ||
| `npm run dev` |
There was a problem hiding this comment.
README contains AI Studio template boilerplate inconsistent with actual app.
The README references:
- Google AI Studio app URL (line 9) — external service unrelated to this monorepo blog app
.env.localsetup (line 18) — the repo uses.env.example, and the app doesn't actually call the Gemini API- AI Studio banner image (line 2) — branding mismatch with Vectorless
This appears to be leftover template content from an AI Studio export. Replace with documentation relevant to the Vectorless blog app, or clarify that this is a demo/template.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@apps/blogs/README.md` around lines 1 - 20, Replace the AI Studio template
boilerplate content in the README with documentation specific to the Vectorless
blog app. Remove the Google AI Studio app URL reference (line 9), update the
environment setup instructions to reference .env.example instead of .env.local
(line 18) and clarify that the app does not require a Gemini API key, and
replace the AI Studio banner image (line 2) with appropriate Vectorless branding
to match the actual application purpose and technology stack.
Comment on lines
+48
to
+50
| title="API Reference" | ||
| href="/docs/api" | ||
| description="The HTTP API surface — generated from the engine's OpenAPI spec." |
There was a problem hiding this comment.
Update the API Reference copy.
This card still says the API reference is "generated from the engine's OpenAPI spec", but this PR removes packages/openapi. That wording is now stale and will mislead readers about the source of truth.
✏️ Suggested copy change
- description="The HTTP API surface — generated from the engine's OpenAPI spec."
+ description="The HTTP API surface — documented in the API reference."🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@apps/docs/content/docs/index.mdx` around lines 48 - 50, The description
property in the API Reference card (title="API Reference" with href="/docs/api")
currently states the API reference is "generated from the engine's OpenAPI
spec", but since this PR removes the packages/openapi package, this description
is now outdated and misleading. Update the description text to accurately
reflect the current source and purpose of the API reference without referencing
the OpenAPI spec generation.
| <div className="flex gap-8"> | ||
| <Link href="/docs" className="text-sm text-vl-muted hover:text-vl-ink">Documentation</Link> | ||
| <Link href="https://github.com/hallelx2/vectorless-engine" className="text-sm text-vl-muted hover:text-vl-ink">GitHub</Link> | ||
| <Link href="#" className="text-sm text-vl-muted hover:text-vl-ink">Status</Link> |
There was a problem hiding this comment.
Replace the placeholder status link with a real destination or non-link text.
Line 292 uses href="#", which creates a dead navigation target in production docs.
Suggested fix
- <Link href="#" className="text-sm text-vl-muted hover:text-vl-ink">Status</Link>
+ <Link href="https://status.vectorless.dev" className="text-sm text-vl-muted hover:text-vl-ink">Status</Link>🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@apps/docs/src/app/`(home)/page.tsx at line 292, The Link component with the
text "Status" currently uses href="#" as a placeholder, which creates a dead
navigation link in production. Either replace the href="#" with a real
destination URL pointing to a valid status page, or if this should not be a
navigation link at all, convert it from a Link component to regular HTML text
wrapped in a span or div element while preserving the existing className
styling.
| @import 'fumadocs-ui/css/neutral.css'; | ||
| @import 'fumadocs-ui/css/preset.css'; | ||
|
|
||
| @source '../**/*.{ts,tsx}'; |
There was a problem hiding this comment.
Include docs content files in Tailwind source scanning.
Line 6 excludes content/**/*.mdx, so utility classes authored in documentation pages can be purged from output.
Suggested fix
`@source` '../**/*.{ts,tsx}';
+@source '../../content/**/*.{md,mdx}';📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| @source '../**/*.{ts,tsx}'; | |
| `@source` '../**/*.{ts,tsx}'; | |
| `@source` '../../content/**/*.{md,mdx}'; |
🧰 Tools
🪛 Stylelint (17.13.0)
[error] 6-6: Unexpected unknown at-rule "@source" (scss/at-rule-no-unknown)
(scss/at-rule-no-unknown)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@apps/docs/src/app/global.css` at line 6, The Tailwind source scanning
configuration in global.css only targets ts and tsx files, which excludes
content/**.mdx documentation files from being scanned for utility classes,
causing those classes to be purged from the output. Modify the `@source` directive
to include the content directory and mdx files by updating the pattern to also
scan documentation markdown files alongside the existing TypeScript sources.
Comment on lines
+21
to
+24
| theme={{ | ||
| enabled: false, | ||
| forcedTheme: 'light', | ||
| }} |
There was a problem hiding this comment.
Enable theme switching if dark mode is intended to remain optional.
Line 22-23 hard-disable theming (enabled: false, forcedTheme: 'light'), which prevents any dark-theme toggle behavior.
Suggested fix
<RootProvider
theme={{
- enabled: false,
- forcedTheme: 'light',
+ enabled: true,
+ defaultTheme: 'light',
}}
>📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| theme={{ | |
| enabled: false, | |
| forcedTheme: 'light', | |
| }} | |
| theme={{ | |
| enabled: true, | |
| defaultTheme: 'light', | |
| }} |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@apps/docs/src/app/layout.tsx` around lines 21 - 24, The theme configuration
object in the layout component has theming disabled with enabled set to false
and forcedTheme set to light, which prevents users from switching between light
and dark themes. To enable theme switching functionality, change the enabled
property to true and adjust the forcedTheme setting appropriately to allow the
theme toggle to work. This will restore the ability for users to switch themes
if dark mode is intended as an optional choice.
Comment on lines
+6
to
+8
| export function GET() { | ||
| return new Response(llms(source).index()); | ||
| } |
There was a problem hiding this comment.
Set an explicit plain-text response header.
llms.txt is returned as a bare string, so the response currently has no Content-Type. That makes the endpoint less reliable for clients that fetch it as a plain-text artifact.
Suggested fix
export function GET() {
- return new Response(llms(source).index());
+ return new Response(llms(source).index(), {
+ headers: {
+ 'Content-Type': 'text/plain; charset=utf-8',
+ },
+ });
}📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| export function GET() { | |
| return new Response(llms(source).index()); | |
| } | |
| export function GET() { | |
| return new Response(llms(source).index(), { | |
| headers: { | |
| 'Content-Type': 'text/plain; charset=utf-8', | |
| }, | |
| }); | |
| } |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@apps/docs/src/app/llms.txt/route.ts` around lines 6 - 8, The GET function
returns a Response without an explicit Content-Type header, which causes the
endpoint to be unreliable for clients expecting plain-text. Modify the Response
constructor in the GET function to include a headers object with the
Content-Type set to "text/plain" so that clients can properly identify the
response format.
Comment on lines
+9
to
+27
| export async function GET(_req: Request, { params }: RouteContext<'/og/docs/[...slug]'>) { | ||
| const { slug } = await params; | ||
| const page = source.getPage(slug.slice(0, -1)); | ||
| if (!page) notFound(); | ||
|
|
||
| return new ImageResponse( | ||
| <DefaultImage title={page.data.title} description={page.data.description} site={appName} />, | ||
| { | ||
| width: 1200, | ||
| height: 630, | ||
| }, | ||
| ); | ||
| } | ||
|
|
||
| export function generateStaticParams() { | ||
| return source.getPages().map((page) => ({ | ||
| lang: page.locale, | ||
| slug: getPageImage(page).segments, | ||
| })); |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
set -euo pipefail
sed -n '1,260p' apps/docs/src/lib/source.ts
sed -n '1,220p' apps/docs/src/app/og/docs/[...slug]/route.tsx
rg -n "getPageImage\(|source\.getPage\(|lang:" apps/docs/src -C 3Repository: hallelx2/vectorless
Length of output: 6141
🏁 Script executed:
# Check the actual directory structure for the route
find apps/docs/src/app -type d -name "*og*" -o -name "*docs*" | sort
# Check for any parent route groups or configurations
rg -n "lang|locale" apps/docs/src/app/og -C 2
# Check the main docs route for comparison
cat -n apps/docs/src/app/docs/[[...slug]]/page.tsx | head -50Repository: hallelx2/vectorless
Length of output: 2830
Remove lang from generateStaticParams() return object.
The route pattern /og/docs/[...slug] has no [lang] segment, yet generateStaticParams() returns lang: page.locale (line 25). This breaks the static params contract—Next.js cannot route to a segment that doesn't exist in the pattern. The lang property should be removed since it's unused by this route.
Additionally, slug.slice(0, -1) (line 11) compensates for getPageImage(page).segments appending 'image.png', indicating a deeper contract mismatch between the helper and route expectations that should be clarified or fixed.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@apps/docs/src/app/og/docs/`[...slug]/route.tsx around lines 9 - 27, In the
generateStaticParams() function, remove the lang: page.locale property from the
returned object since the route pattern /og/docs/[...slug] does not have a
[lang] segment. The function should only return slug:
getPageImage(page).segments. Additionally, investigate why slug.slice(0, -1) is
needed in the GET function while getPageImage(page).segments appends
'image.png', and either clarify or fix this contract mismatch between the helper
function and route expectations.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR consolidates the vectorless-docs repository and the new vectorless-blog repository into the main web monorepo under apps/docs and apps/blogs. Deprecated/obsolete packages (ts-sdk, openapi, mcp-tools) have been removed, database migrations applied to Neon, and Next.js builds verified.
Summary by Sourcery
Consolidate the standalone docs and blog properties into the main web monorepo by adding a new Fumadocs-powered docs app and a marketing-style blog app, while removing deprecated SDK, OpenAPI, and MCP tools packages.
New Features:
apps/docsNext.js application using Fumadocs for Vectorless documentation, including home, concepts, getting-started, SDK, API, and self-hosting content skeletons.apps/blogsNext.js application providing a Vectorless marketing blog with featured articles and an interactive retrieval demo UI.Enhancements:
Chores:
ts-sdk,openapi,mcp-tools) and associated config from the monorepo to simplify the codebase.Summary by CodeRabbit
Release Notes
New Features
Chores