mirror of
https://github.com/open-webui/open-webui.git
synced 2026-05-06 10:58:17 -05:00
[PR #22347] [CLOSED] perf: Add in-memory cache for MCP tool specs #49671
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
📋 Pull Request Information
Original PR: https://github.com/open-webui/open-webui/pull/22347
Author: @shshzi
Created: 3/7/2026
Status: ❌ Closed
Base:
dev← Head:main📝 Commits (10+)
fe6783cMerge pull request #19030 from open-webui/devfc05e0aMerge pull request #19405 from open-webui/deve3faec6Merge pull request #19416 from open-webui/dev9899293Merge pull request #19448 from open-webui/dev140605eMerge pull request #19462 from open-webui/dev6f1486fMerge pull request #19466 from open-webui/devd95f533Merge pull request #19729 from open-webui/deva7271530.6.43 (#20093)6adde20Merge pull request #20394 from open-webui/devf9b0534Merge pull request #20522 from open-webui/dev📊 Changes
6 files changed (+98 additions, -13 deletions)
View changed files
📝
backend/open_webui/env.py(+9 -0)📝
backend/open_webui/main.py(+4 -2)📝
backend/open_webui/routers/configs.py(+3 -0)➕
backend/open_webui/utils/mcp/cache.py(+44 -0)📝
backend/open_webui/utils/mcp/client.py(+3 -1)📝
backend/open_webui/utils/middleware.py(+35 -10)📄 Description
Pull Request Checklist
Note to first-time contributors: Please open a discussion post in Discussions to discuss your idea/fix with the community before creating a pull request, and describe your changes before submitting a pull request.
This is to ensure large feature PRs are discussed with the community first, before starting work on it. If the community does not want this feature or it is not relevant for Open WebUI as a project, it can be identified in the discussion before working on the feature and submitting the PR.
Before submitting, make sure you've checked the following:
devbranch. PRs targetingmainwill be immediately closed.devto ensure no unrelated commits (e.g. frommain) are included. Push updates to the existing PR branch instead of closing and reopening.Changelog Entry
Description
Add in-memory cache for MCP tool specs to eliminate redundant network round-trips during chat completions.
Currently, every chat completion with MCP tools connects to each MCP server and fetches the full tool list, even though specs rarely change. This adds an in-memory cache (72-hour default TTL, configurable via MCP_TOOL_SPECS_CACHE_TTL) that skips both the connection and tool listing on cache hit - only connecting to the MCP server if the model actually calls a tool. Cache keys are scoped system-wide for static auth (bearer/none) and per-user for credential-forwarding auth types. Cache invalidates automatically on tool server config changes.
Also fixes MCPClient.disconnect() to safely handle unconnected instances.
Added
Changed
Deprecated
Removed
Fixed
Security
Breaking Changes
Additional Information
Screenshots or Videos
Contributor License Agreement
By submitting this pull request, I confirm that I have read and fully agree to the Contributor License Agreement (CLA), and I am providing my contributions under its terms.
🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.