Skip to content

docs: clarify API server tool execution locality - #19117

Closed
pingchesu wants to merge 1 commit into
NousResearch:mainfrom
pingchesu:feat/remote-agent-local-tools
Closed

docs: clarify API server tool execution locality#19117
pingchesu wants to merge 1 commit into
NousResearch:mainfrom
pingchesu:feat/remote-agent-local-tools

Conversation

@pingchesu

Copy link
Copy Markdown
Contributor

Summary

First small PR for #18715.

  • Clarifies that Hermes' OpenAI-compatible API server is a server-side Hermes agent runtime, not a pure LLM proxy
  • Documents that tools run on the API-server host in current mode
  • Adds /v1/capabilities.runtime metadata so clients can discover current server_agent / server-side tool execution semantics

Why

When a local client points model.base_url at a remote Hermes API server, users may expect normal LLM-provider semantics where tool calls are returned to the client and executed locally. Today the API server creates its own AIAgent, so tools execute on the API-server host. This PR documents and advertises that behavior before split-runtime work.

Test Plan

  • python -m pytest tests/gateway/test_api_server.py::TestCapabilitiesEndpoint -q
  • python -m pytest tests/gateway/test_api_server.py -q

Related

Part 1 of the staged work for #18715. Follow-up PRs can add the tool locality/executor abstraction and then the explicit split-runtime bridge.

@alt-glitch alt-glitch added type/docs Documentation improvements P3 Low — cosmetic, nice to have comp/gateway Gateway runner, session dispatch, delivery labels May 3, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

comp/gateway Gateway runner, session dispatch, delivery P3 Low — cosmetic, nice to have type/docs Documentation improvements

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants