Your agent stays your choice.

Your team’s context and plans are files in your git, not settings inside an agent. Whichever agent you run in a worktree finds them there. Claude Code and Codex are built in today.

Apply for the private beta Read the agents docs

Context lands for whichever agent you run

CLAUDE.md, AGENTS.md, .claude/skills/, .codex/, or any other instruction file or doc: whatever an agent’s context names lands as a file in its worktree, so any agent that reads its worktree gets it.

Files land as you wrote them. CodeSense doesn’t turn CLAUDE.md into AGENTS.md, so a team that uses both keeps both, or points one at the other.

How context lands in a worktree
  • ledgerline-api/ (a worktree) /
    • CLAUDE.md from your team context
    • AGENTS.md from your team context
    • .claude/ /
      • skills/review-pr/SKILL.md from your team context
      • settings.json your repo's own, untouched
      • settings.local.json CodeSense's own hook
    • src/ your code, untouched
One worktree, after CodeSense has written the team's context into it.

Built in today: Claude Code and Codex

Claude Code and Codex are built in, with saved agents, accounts, headless runs and CodeSense’s tools. Live status is Claude-first today.

Launch
Start your agent in a story’s terminal, or launch one with a chosen account into an open story workspace.
MCP tools
CodeSense’s tools are wired into Claude Code and Codex automatically, scoped per launch profile.
Accounts and usage
Sign in and see usage for your Claude Code and Codex accounts.
Headless runs
Headless runs use each agent’s own non-interactive mode.
Agents and runtimes

Switch between Claude Code and Codex without redoing anything

Context and stories are files that no agent owns. Switch between Claude Code and Codex without redoing anything: the next story starts from the same context, the same plan and the same checks.

Switch agents

Models are the agent's business

CodeSense never calls a model itself. Your agent does, with whatever model you’ve configured it to use.

Running another agent? Any CLI agent you’ve installed runs in a worktree terminal and reads the same context files. It doesn’t get launch profiles, live status or accounts yet.

Run your own tools in a worktree

Coming

Pick the agent per story. The context, the plan and the proof stay put.

Apply for the private beta