Add peer/custom LLM integrations to roadmap

Author: philz42 Channel: #feedback Link: https://discord.com/channels/1405685085923049482/1405686161791516873/1549199403024257026

Summary

Add support for "peer integrations" as "custom LLM integrations" to the product roadmap.

Context

This feature request stems from a need to allow LLM integrations to access private machines (like exe.dev machines) without requiring them to be publicly exposed.

Related Use Case

Users currently need to set up public proxies to route LLM integration requests to private machines. For example, a proxy setup like opencode-go-shelley-shell-open-go-code requires the host machine to be public, which is a security/privacy concern.

Additional Context

A related request mentioned the need for multiple OpenAI integrations to work properly, as some models don't work when authenticated with the subscription (though endpoints may be a workaround).

Acceptance Criteria

• Evaluate architecture for peer/custom LLM integrations
• Add to product roadmap with estimated timeline
• Ensure integration can access private machines securely