Skip to content

Name Unbiased AI in the model list and the docs index - #1422

Open
badjer wants to merge 1 commit into
CopilotKit:mainfrom
circuitandchisel:docs/unbiased-ai-model-list
Open

badjer wants to merge 1 commit into
CopilotKit:mainfrom
circuitandchisel:docs/unbiased-ai-model-list

Conversation

@badjer

@badjer badjer commented Nov 10, 2026

Copy link
Copy Markdown
Contributor

What this changes

Two documentation lines, following up #1067 (Sign in with Unbiased AI, shipped in 0.3.3):

  • docs/configuration.md, "The model list": a row for an Unbiased AI key or sign-in. Every other sign-in the desktop offers has one. The row states what the server does with that credential today: the key is written as OPENAI_BASE_URL=https://api.unbiased.ai/v1, modelSourceFor classifies that host as a gateway, so the listing is the gateway's own /models as it is, and because parseCapabilities reads models.dev for the four vendors a deployment can name directly and none of them lists Pareto, the model runs on its own default effort ("a model nothing describes yet is still listed").
  • docs/README.md: the index line for Desktop provider OAuth said that doc covers Google and xAI sign-in. It has covered Unbiased AI too since 0.3.3.

Nothing is added to shared/model-providers.json: that file names API families (openai, anthropic, google) that every harness constructs a client for, and Unbiased AI's gateway is reached as openai with a base URL, exactly as xAI's key is. The row documents that, rather than inventing a provider id.

Where it runs

Documentation only. No code, state, listener or schedule changes.

  • New state that outlives a request? None.
  • What happens on the second replica? Nothing changes; no code is touched.
  • Anything serialised? No.
  • Anything fanned out to a browser? No.
  • New listener, port, or schedule? No.

Boundary and audit

  • Every acting call still goes through the gateway: no acting call is touched.
  • New refusals and new failures each write a row: none added.
  • Nothing new is trusted from the client: nothing added.

Changelog

  • No line: a deployment behaves no differently afterwards. Only two lines of documentation change.

Proof

  • The table row was written from the code it describes: server/src/model-catalogue.ts (modelSourceFor returns "gateway" for api.unbiased.ai; listingRequest reads ${OPENAI_BASE_URL}/models for a gateway; CAPABILITY_VENDORS is openai, anthropic, google, xai).
  • models.dev's api.json has no Pareto model under any of those four vendors (its pareto provider is a different company, Pareto Inference; Unbiased's models appear only namespaced under OpenRouter), so the third column is accurate.
  • Rendered the two Markdown files locally; the table keeps its three columns.

🤖 Generated with Claude Code

https://claude.ai/code/session_012nVJiRyd2HT6qtKHEjt9bN

@badjer
badjer force-pushed the docs/unbiased-ai-model-list branch 2 times, most recently from 97c3556 to 3f763a1 Compare November 10, 2026 15:55
The model list table says where each sign-in's models and thinking levels
come from, and named every sign-in the desktop offers except Unbiased AI,
which arrived in 0.3.3. Its row says what the server does with that key:
the list is read from the gateway's own `/models`, and no vendor models.dev
describes names Pareto, so it runs on its own default effort.

Documentation only: no deployment behaves differently afterwards.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012nVJiRyd2HT6qtKHEjt9bN
@badjer
badjer force-pushed the docs/unbiased-ai-model-list branch from 3f763a1 to b8f012a Compare November 11, 2026 03:57

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