chore: Retire openrag-mcp; switch docs to streamable HTTP#1668
Merged
Conversation
Remove the stdio-based MCP server and all in-repo MCP tooling, and update README to mark the package as retired. Deleted module files include the MCP entrypoint, server, config, registry and individual tools (chat, search, documents, settings). The README was rewritten to announce that openrag-mcp is retired, explain migration to the built-in streamable-HTTP /mcp endpoint, update Cursor/Claude examples to use URL+headers auth, list the new v1 API tools, and note that the last PyPI release is final. This change consolidates MCP functionality into the OpenRAG core and removes the subprocess/stdio implementation and its source code.
Contributor
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Run ID: ⛔ Files ignored due to path filters (1)
📒 Files selected for processing (11)
💤 Files with no reviewable changes (9)
WalkthroughThis PR deprecates the standalone ChangesOpenRAG MCP SDK Deprecation
🎯 2 (Simple) | ⏱️ ~10 minutes
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 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 |
Update package metadata to reflect retirement and integration into the OpenRAG backend. Bump version to 0.3.0 and replace the project description with a retirement/migration note. Set Development Status to Inactive, remove explicit Python version classifiers, and clear runtime dependencies and the CLI script entrypoint. Also remove the hatch env pip-args setting; build-system and wheel package target remain unchanged.
Contributor
There was a problem hiding this comment.
Pull request overview
Retires the standalone openrag-mcp (stdio/subprocess) implementation by deleting the in-repo MCP server/tooling and rewriting the SDK README to direct users to OpenRAG’s built-in streamable-HTTP MCP endpoint (/mcp).
Changes:
- Deleted the Python MCP server entrypoint, config, and all tool implementations under
sdks/mcp/src/openrag_mcp/. - Rewrote
sdks/mcp/README.mdto announce retirement and provide Cursor/Claude “url + headers” connection examples to/mcp. - Documented the built-in MCP surface area (chat/search/settings/docs/tasks/knowledge-filters) in the README.
Reviewed changes
Copilot reviewed 11 out of 12 changed files in this pull request and generated 4 comments.
Show a summary per file
| File | Description |
|---|---|
| sdks/mcp/src/openrag_mcp/tools/settings.py | Removed settings/models tools implementation (deleted). |
| sdks/mcp/src/openrag_mcp/tools/search.py | Removed search tool implementation (deleted). |
| sdks/mcp/src/openrag_mcp/tools/registry.py | Removed MCP tool registry (deleted). |
| sdks/mcp/src/openrag_mcp/tools/documents.py | Removed document/task tool implementations (deleted). |
| sdks/mcp/src/openrag_mcp/tools/chat.py | Removed chat tool implementation (deleted). |
| sdks/mcp/src/openrag_mcp/tools/init.py | Removed tool registration module (deleted). |
| sdks/mcp/src/openrag_mcp/server.py | Removed stdio MCP server entrypoint (deleted). |
| sdks/mcp/src/openrag_mcp/config.py | Removed env/config + client wiring (deleted). |
| sdks/mcp/src/openrag_mcp/main.py | Removed module entrypoint (deleted). |
| sdks/mcp/README.md | Updated documentation to retirement notice + migration to /mcp streamable HTTP. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| > | ||
| > OpenRAG now ships a built-in [Model Context Protocol](https://modelcontextprotocol.io/) server over **streamable HTTP** — no subprocess, no separate install, no API key in a subprocess env block. Connect your MCP client directly to your running OpenRAG instance. | ||
| > | ||
| > The last published version of `openrag-mcp` on PyPI is the final release. No further updates will be made to this package. |
|
|
||
| Document ingestion and management tools (`openrag_ingest_file`, `openrag_ingest_url`, `openrag_delete_document`, `openrag_get_task_status`, `openrag_wait_for_task`) are implemented but not yet registered in this server; they will be enabled in a future release. | ||
| - A running OpenRAG instance (v0.3.0 or later) | ||
| - An OpenRAG API key — create one in **Settings → API Keys** |
| | `openrag_get_chat` | Get a specific chat conversation by ID. | | ||
| | `openrag_delete_chat` | Delete a chat conversation by ID. | | ||
| | `openrag_search` | Semantic search over the knowledge base. Supports filters, score threshold, data sources. | | ||
| | `openrag_ingest` | Ingest documents (files, URLs, text) into the knowledge base. Returns a `task_id`. | |
Comment on lines
+101
to
104
| ## Available tools | ||
|
|
||
| ## Example prompts | ||
| All tools are auto-exposed from the `/v1/` API and are available immediately after connecting: | ||
|
|
phact
approved these changes
May 22, 2026
ricofurtado
pushed a commit
that referenced
this pull request
May 23, 2026
* Retire openrag-mcp; switch docs to streamable HTTP Remove the stdio-based MCP server and all in-repo MCP tooling, and update README to mark the package as retired. Deleted module files include the MCP entrypoint, server, config, registry and individual tools (chat, search, documents, settings). The README was rewritten to announce that openrag-mcp is retired, explain migration to the built-in streamable-HTTP /mcp endpoint, update Cursor/Claude examples to use URL+headers auth, list the new v1 API tools, and note that the last PyPI release is final. This change consolidates MCP functionality into the OpenRAG core and removes the subprocess/stdio implementation and its source code. * Mark MCP SDK retired and clean package metadata Update package metadata to reflect retirement and integration into the OpenRAG backend. Bump version to 0.3.0 and replace the project description with a retirement/migration note. Set Development Status to Inactive, remove explicit Python version classifiers, and clear runtime dependencies and the CLI script entrypoint. Also remove the hatch env pip-args setting; build-system and wheel package target remain unchanged. * chore: update uv.lock files after version bump * Update uv.lock --------- Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
ricofurtado
pushed a commit
that referenced
this pull request
May 23, 2026
* Retire openrag-mcp; switch docs to streamable HTTP Remove the stdio-based MCP server and all in-repo MCP tooling, and update README to mark the package as retired. Deleted module files include the MCP entrypoint, server, config, registry and individual tools (chat, search, documents, settings). The README was rewritten to announce that openrag-mcp is retired, explain migration to the built-in streamable-HTTP /mcp endpoint, update Cursor/Claude examples to use URL+headers auth, list the new v1 API tools, and note that the last PyPI release is final. This change consolidates MCP functionality into the OpenRAG core and removes the subprocess/stdio implementation and its source code. * Mark MCP SDK retired and clean package metadata Update package metadata to reflect retirement and integration into the OpenRAG backend. Bump version to 0.3.0 and replace the project description with a retirement/migration note. Set Development Status to Inactive, remove explicit Python version classifiers, and clear runtime dependencies and the CLI script entrypoint. Also remove the hatch env pip-args setting; build-system and wheel package target remain unchanged. * chore: update uv.lock files after version bump * Update uv.lock --------- Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
ricofurtado
added a commit
that referenced
this pull request
May 26, 2026
* fix: Ensure SUCCESS status requires fetchable result in DoclingPollingService * style: ruff autofix (auto) * fix: Catch specific DoclingServeError when fetching task result after SUCCESS status * feat: update style for oss of the failed task in the task panel (#1647) * update style for oss of the failed task in the task panel * keep logic on click, remove unecessary useeffect * fix padding * wip implementing Saas style * utils to reshape error until backend provide info we need * utils to reshape error until backend provide info we need * utils to reshape error until backend provide info we need and fixinf fallbacks of isTotalFailure * utils to reshape error until backend provide into * have Saas style for failed and complete labelstatus and width and border * few style adjustment to follow codebase pattern * adjust succeed and partially succeed case * adding comment for TODO implementation or more clarity * remove carbon icon package and replace carbon icon * add incident-reporter-icon --------- Co-authored-by: Olfa Maslah <[email protected]> * fix: Encode IBM API key as Basic auth header (#1664) * Encode IBM API key as Basic auth header Add base64 encoding for the IBM auth path: import base64, construct a Basic auth token from X-Username and X-Api-Key (username:apikey), and store it in user.jwt_token and user.opensearch_credentials. Also set request.state.user before attaching the DB user ID so downstream code can access the created user object. * style: ruff autofix (auto) --------- Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com> * fix: restart deployment if env changes (#1665) * restart deployment if env changes * unit test * lint * fix: Ensure Langflow .env variable definitions from LANGFLOW_VARIABLES_TO_GET_FROM_ENVIRONMENT (#1667) * Ensure we dynamically update the list of Langflow .env environment variables with default values when the comma separated list defined in LANGFLOW_VARIABLES_TO_GET_FROM_ENVIRONMENT changes * fix tests * fix additional linting errors --------- Co-authored-by: rodageve <[email protected]> * chore: Retire openrag-mcp; switch docs to streamable HTTP (#1668) * Retire openrag-mcp; switch docs to streamable HTTP Remove the stdio-based MCP server and all in-repo MCP tooling, and update README to mark the package as retired. Deleted module files include the MCP entrypoint, server, config, registry and individual tools (chat, search, documents, settings). The README was rewritten to announce that openrag-mcp is retired, explain migration to the built-in streamable-HTTP /mcp endpoint, update Cursor/Claude examples to use URL+headers auth, list the new v1 API tools, and note that the last PyPI release is final. This change consolidates MCP functionality into the OpenRAG core and removes the subprocess/stdio implementation and its source code. * Mark MCP SDK retired and clean package metadata Update package metadata to reflect retirement and integration into the OpenRAG backend. Bump version to 0.3.0 and replace the project description with a retirement/migration note. Set Development Status to Inactive, remove explicit Python version classifiers, and clear runtime dependencies and the CLI script entrypoint. Also remove the hatch env pip-args setting; build-system and wheel package target remain unchanged. * chore: update uv.lock files after version bump * Update uv.lock --------- Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com> * fix: connector sync and override feature (#1663) * Add filename-based duplicate handling for connectors Add end-to-end support for filename-based duplicate handling on connector ingests. Frontend: send a new replace_duplicates flag with connector sync requests, perform a pre-sync duplicate check, and show a DuplicateHandlingDialog that lets users overwrite or skip duplicates when uploading from provider UI. Backend: propagate replace_duplicates through connector_router, request models, and connector services into the file processors. ConnectorFileProcessor and LangflowConnectorFileProcessor now check whether a filename already exists in the index and either fail the file task or delete the existing document before ingesting when replace_duplicates is true. Utilities/tests: clean_connector_filename now preserves original spacing/slashes and only enforces MIME-mapped extensions; get_filename_aliases adds underscore/sanitized variants so lookups match connector-indexed names. Add unit tests covering filename dedupe logic and filename alias behavior. * Use duplicateNames list and display names Replace numeric duplicateCount with a duplicateNames string[] across upload and dropdown flows so the UI can show the actual file names that would be overwritten. The duplicate-handling dialog now accepts duplicateNames, derives an effective count, and lists up to 5 duplicate filenames with an "… and N more" indicator; message labels and button text use the effective count. Toast messages and pending state in upload/[provider]/page.tsx and knowledge-dropdown.tsx were updated to pass and consume duplicateNames and to use duplicateNames.length for counts. * Update page.tsx * style: ruff autofix (auto) --------- Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com> * fix: update OAuth prompt to consent for connector mutation (#1657) * fix: implement transient error handling for Docling result fetch * style: ruff autofix (auto) * refactor: remove unused import of Optional in docling_polling_service.py * refactor: change PollOutcome to use StrEnum for better type safety * refactor: enhance task status endpoints with structured failure metadata * style: ruff autofix (auto) * revert "style: ruff autofix (auto)" This reverts commit bc8be33. * style: ruff autofix (auto) * fix: Ensure SUCCESS status requires fetchable result in DoclingPollingService * style: ruff autofix (auto) * fix: Catch specific DoclingServeError when fetching task result after SUCCESS status * fix: implement transient error handling for Docling result fetch * style: ruff autofix (auto) * refactor: remove unused import of Optional in docling_polling_service.py * refactor: change PollOutcome to use StrEnum for better type safety * refactor: enhance task status endpoints with structured failure metadata * style: ruff autofix (auto) * revert "style: ruff autofix (auto)" This reverts commit bc8be33. * style: ruff autofix (auto) * Update tests/unit/test_task_service_get_task_status2.py Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> * style: ruff autofix (auto) * fix: handle timeout during Docling result fetch after SUCCESS status * fix: update task status checks to use enum values for consistency * fix: enhance failure metadata for duplicate file errors in ingestion * style: ruff autofix (auto) --------- Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com> Co-authored-by: Wallgau <[email protected]> Co-authored-by: Olfa Maslah <[email protected]> Co-authored-by: Edwin Jose <[email protected]> Co-authored-by: ming <[email protected]> Co-authored-by: rodageve <[email protected]> Co-authored-by: rodageve <[email protected]> Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com> Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
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.
Remove the stdio-based MCP server and all in-repo MCP tooling, and update README to mark the package as retired. Deleted module files include the MCP entrypoint, server, config, registry and individual tools (chat, search, documents, settings). The README was rewritten to announce that openrag-mcp is retired, explain migration to the built-in streamable-HTTP /mcp endpoint, update Cursor/Claude examples to use URL+headers auth, list the new v1 API tools, and note that the last PyPI release is final. This change consolidates MCP functionality into the OpenRAG core and removes the subprocess/stdio implementation and its source code.
Summary by CodeRabbit
Chores
openrag-mcppackage. Users should now connect to OpenRAG's built-in MCP server via HTTP at the/mcpendpoint instead.Documentation