Agent credentials
Configure API keys and connected plans for coding agents.
Superconductor primarily works on a BYOK (bring-your-own-keys) basis, meaning that coding agents use your own keys or connected plans. If your workspace has credits, supported agents can also use those credits when no higher-priority credential is available.
Connecting credentials
Agent API keys can be defined on the Workspace, or on the Project.
For Claude Code, Codex, and Grok Build agents, we additionally support individual users connecting their Claude, ChatGPT, and SuperGrok premium plans instead of using API keys. This can be done either in Workspace settings, Project settings, or Account settings.
See the Usage dashboard to track agent spend across your workspace and spot metered-API spend a connected plan could cover.
How do I track usage? If an implementation uses an API key or workspace credits, the implementation card's Cost value estimates the API spend for that run. You can track remaining credits under Workspace Settings → Billing. If an implementation uses a connected Claude or ChatGPT plan, use the 5-hour and 7-day usage stats shown next to the agent name instead. The card's Cost value may still appear, but it is an API-price estimate based on the agent's reported token usage, not a measure of charges or credits consumed on your plan.
How are Gemini API keys used? Gemini API keys configured in Workspace or Project Settings are used only for Gemini-powered Superconductor skills, such as Video Understanding and Image Generation. To run coding agents and/or benchmarks with Gemini models, use OpenCode, Factory Droid, or Cursor.
Deciding which credentials to use
If an implementation agent is launched through an integration (e.g. by an external PR, Slack, Email triage, or the Meeting assistant), Superconductor can use the associated user's connected plan when one is available for the selected agent. Otherwise, we use Project API keys, Workspace API keys, and then workspace credits when they are available for the selected agent.
If the agent is launched manually by a user, we follow this precedence order:
- That user's connected plan (if any)
- Project API key (if any)
- Workspace API key
- Workspace credits, if available for the selected agent
- Another workspace member's connected plan, for benchmark runs
What if my connected plan hits its usage limit? Superconductor does not automatically fall back on an API key mid-run. The implementation will stop, and you'll see a message in the chat indicating that your plan's limit has been reached and when it resets. For future runs, you can disconnect your plan in your settings and use an API key or credits instead until the limit resets.
Best available credentials
Some features don't ask you which agent to use — they automatically run with the best coding agent you have credentials configured for. Superconductor tries them in this order and uses the first one with a working credential:
- Codex (GPT-5.5, high reasoning)
- Claude Code (Opus 4.8)
- Amp
- OpenCode (GLM 5.2)
- Pi (GLM 5.2)
- Grok Build (high reasoning)
- Grok Composer 2.5 (high reasoning)
- Cursor (Composer 2.5)
Features that use this fallback:
- Automatic Environment Setup
- Guided review
- QA Checks [Beta]
- Email triage [Beta]
- Meeting assistant [Beta]
- The fallback default selection in the profile selector when you launch an implementation, if your Default agent profile is unset or unavailable for the project
Benchmarking
Agent API keys are required to include agents in Benchmark runs. Only agents with configured keys or available credits will be available for selection when setting up a benchmark.