How do you use Claude Code models in OpenCode without new API keys?
Paying for Claude Code but forced into a second auth setup inside OpenCode is avoidable friction. The new built-in Claude Agent provider reuses your Claude Code login and models with a one-time connect step.
What it is
A built-in provider in OpenCode that wraps your existing Claude Code installation. You install Claude Code, authenticate once, and select Claude Agent from /connect.
- OPENCODE_CLAUDE_PATH overrides the Claude executable if it is not on your default path
Applied case
A developer on a Claude Code subscription wants OpenCode's desktop handling for long sessions. They install Claude Code and run claude auth login. Then they pick Claude Agent in /connect, and their subscription carries over with no new API key. The same release cut rendering work for streaming replies, so models stay responsive in extended conversations.
- Reuse Claude Code auth in OpenCode via /connect, no new key required
Limitations
This does not give you Claude access without a Claude Code install and login. Managed policies also changed: a bare integration.use rule no longer covers MCP servers, skills, or plugins, so review your rules before upgrading.
When to use it
Use when you already have a Claude Code login and want it inside OpenCode. Skip when you rely on other providers or cannot install Claude Code locally.
Read more
Recommended for you
- WorkflowOpencode
How do you keep coding agents on task without babysitting every step?
Agent sessions drift from the objective, forcing you to watch every turn; opencode's /goal lets a session run until a reviewer confirms the goal is met.
- WorkflowHerdr
How do you supervise coding agents without watching every window?
I run Claude and opencode side by side in one terminal. Three herdr habits replaced window-polling with state-waiting: persistent sessions, the agent loop, a setup script.
- WorkflowSuperpowersClaude Code
Why does your multi-agent workflow keep colliding?
Two agents in two threads share files but not context. Both decide on stale state. Fix: one fresh agent per task, with isolated context.
- WorkflowSuperpowersClaude Code
What makes agent output trustworthy when the model is not the problem?
The reliability problem is not the model. It is the missing disciplines around it. Brainstorm, plan, test, verify, review: each gate before the next step.
Enjoyed this article?
Subscribe for new articles. No spam. Unsubscribe anytime.