Skip to content

Add keyless Parallel Search MCP to OpenWorld - #51

Open
georgeatparallel wants to merge 1 commit into
modelscope:mainfrom
georgeatparallel:parallel-search-2a03d0d4-171b-4f13-ad6c-09f1d52516d4
Open

georgeatparallel wants to merge 1 commit into
modelscope:mainfrom
georgeatparallel:parallel-search-2a03d0d4-171b-4f13-ad6c-09f1d52516d4

Conversation

@georgeatparallel

Copy link
Copy Markdown

Description

Give OpenWorld agents free web search and page fetching with Parallel Search MCP. Anonymous search uses Fast mode, with no account or API key. Unconfigured installations use Parallel; explicit server files, edits to the legacy finance configuration, and empty server maps stay selected.

Keep MCP connections on one owning task across discovery, calls and cleanup. Each environment owns its research session ID, including retries and new-conversation resets.

Run python -m unittest discover -s tests -p test_openworld_search.py -v to check the resolver, dispatcher, connection lifecycle and source results. All eight focused tests pass in a clean OpenWorld dependency environment.

I work at Parallel.

Checklist

  • All tests are passing (focused OpenWorld tests pass; full GPU training suites not run)
  • Docstrings are in Google style
  • Related documentation has been updated (e.g. links, examples, etc.)
  • Code is ready for review

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant