feat(custom_providers_support): support custom OpenAI/Anthropic-compatible LLM endpoints - #73
Open
erbanku wants to merge 2 commits into
Open
feat(custom_providers_support): support custom OpenAI/Anthropic-compatible LLM endpoints#73erbanku wants to merge 2 commits into
erbanku wants to merge 2 commits into
Conversation
…tible LLM endpoints Add configurable provider, base URL, and model resolution for semantic search. Enable OpenAI-compatible, Anthropic-compatible, local, and enterprise-hosted endpoints via environment variables.
Author
|
I added myself to the author field in I have published it here: https://www.npmjs.com/package/@erbanku/lat.md |
This was referenced Jul 4, 2026
erbanku
marked this pull request as ready for review
July 4, 2026 03:41
There was a problem hiding this comment.
Pull request overview
This PR extends lat search/lat expand semantic search configuration to support custom OpenAI-compatible and Anthropic-compatible embedding endpoints (base URL, provider header format, model override), and adds runtime dimension resolution so the vectors DB schema can adapt to non-default embedding models.
Changes:
- Add env/config resolvers for
LAT_LLM_BASE_URL,LAT_LLM_PROVIDER,LAT_LLM_MODEL, andLAT_LLM_ANTHROPIC_VERSION, and wire them into provider detection. - Introduce
resolveSchema()to probe/cached embedding dimensions and trigger automatic DB rebuild when provider/model changes. - Expand tests and documentation to cover custom endpoints and dimension resolution behavior.
Reviewed changes
Copilot reviewed 17 out of 17 changed files in this pull request and generated 5 comments.
Show a summary per file
| File | Description |
|---|---|
| tests/search.test.ts | Adds unit tests for custom endpoint provider detection and model override dimension behavior. |
| tests/schema.test.ts | Adds tests for schema dimension probing + caching using a local embeddings stub server. |
| tests/config.test.ts | Adds tests for env/config resolution of custom endpoint settings and bundling into provider options. |
| templates/cursor-rules.md | Updates guidance for semantic search setup to mention custom endpoints. |
| templates/AGENTS.md | Updates agent guidance for semantic search setup to mention custom endpoints. |
| src/search/schema.ts | New schema resolution layer that probes/caches embedding dimensions and detects provider/model changes. |
| src/search/provider.ts | Adds ProviderOptions, custom base URL support, Anthropic-compatible header mode, and model overrides that clear static dims. |
| src/config.ts | Adds config fields + env/config resolution functions for custom endpoint/provider/model settings. |
| src/cli/search.ts | Uses provider options + schema resolution; drops/rebuilds DB table when provider signature changes; improves CLI error handling paths. |
| src/cli/init.ts | Adds CLI init messaging describing advanced custom-endpoint environment variables. |
| src/cli/index.ts | Enhances lat config output to show resolved LLM settings. |
| README.md | Documents custom endpoint configuration and new env vars/config fields. |
| package.json | Updates package metadata (name/author/repo/homepage/publishConfig) alongside the feature changes. |
| lat.md/tests/search.md | Documents new/expanded tests for configuration, provider detection, and dimension resolution. |
| lat.md/cli.md | Documents new config fields, provider detection order, and the dimension resolution mechanism. |
| CLAUDE.md | Updates agent instructions for semantic search setup to mention custom endpoints. |
| AGENTS.md | Updates agent instructions for semantic search setup to mention custom endpoints. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+69
to
+73
| let dimensions = provider.dimensions; | ||
| if (dimensions === undefined && !configChanged) { | ||
| const cached = await getMeta(db, DIMENSIONS_META_KEY); | ||
| if (cached) dimensions = parseInt(cached, 10); | ||
| } |
Comment on lines
+162
to
+166
| try { | ||
| await runIndex(ctx.latDir, key, progress); | ||
| } catch (err) { | ||
| return { output: (err as Error).message, isError: true }; | ||
| } |
Comment on lines
+170
to
+175
| let result: SearchResult; | ||
| try { | ||
| result = await runSearch(ctx.latDir, query, key, opts.limit, progress); | ||
| } catch (err) { | ||
| return { output: (err as Error).message, isError: true }; | ||
| } |
Comment on lines
+2
to
16
| "name": "@erbanku/lat.md", | ||
| "version": "0.11.2", | ||
| "description": "A knowledge graph for your codebase, written in markdown", | ||
| "type": "module", | ||
| "packageManager": "pnpm@10.30.2", | ||
| "license": "MIT", | ||
| "author": "Yury Selivanov", | ||
| "author": "Erban Ku", | ||
| "repository": { | ||
| "type": "git", | ||
| "url": "https://github.com/1st1/lat.md.git" | ||
| "url": "git+https://github.com/erbanku/lat.md.git" | ||
| }, | ||
| "homepage": "https://github.com/erbanku/lat.md", | ||
| "publishConfig": { | ||
| "access": "public" | ||
| }, |
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
Collaborator
|
@erbanku I've just implemented local embeddings support as native a default option for lat.
Anthropic doesn't seem to have embeddings models/api, what does anthropic-compatible mean in this context? |
Author
|
Anthropic-compatible here means the API format (headers and request shape), not Anthropic’s own models. Some third-party providers expose embeddings behind that Anthropic-style API, which is what I wanted to support. Anthropic itself doesn’t have an embeddings API, as you said. Happy to clarify that in the docs if the naming is confusing. |
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
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.
Add configurable provider, base URL, and model resolution for semantic search.
Enable OpenAI-compatible, Anthropic-compatible, local, and enterprise-hosted endpoints via environment variables.
Closes #72