Coding agents
Rius traces the coding agents your team already uses. Each session becomes one trace with its model calls and tool calls, so you can see what an agent did, where it failed, and, for most agents, what it cost. Claude Code, Codex, and Cursor each install the same Rius plugin. OpenHands needs no plugin: its SDK already sends OpenTelemetry, the open standard Rius takes in.
Tracing is off until each developer turns it on: per folder for the plugin, by setting environment variables for OpenHands.
Compare the agents
| Claude Code | Codex (beta) | Cursor (beta) | OpenHands | |
|---|---|---|---|---|
| How you connect it | Rius plugin | Rius plugin | Rius plugin | Three environment variables |
| What one trace covers | A session | A session | An agent chat | A conversation |
| Tokens and cost | ✓ | ✓ | ✗ | ✓, with gaps (see below) |
| How each developer shows on Users | By sign-in email | By sign-in email | By sign-in email | By the user_id you pass |
| Prompts, replies, and tool output | Sent only from folders you opt in | Sent only from folders you opt in | Sent only from folders you opt in | Always sent |
| Query your traces over MCP | From Claude Code | From Codex | From Cursor | From any MCP client |
- Cost is worked out from token counts at API rates. On a subscription plan it measures usage, not what you pay.
- OpenHands has no token data for OpenAI Responses API models or for the web app’s streamed calls, and OpenHands Cloud can’t send to Rius; see its Limits.
- Developers by email needs plugin 0.6.2 or later and the plugin’s own sign-in. Sessions sent with an existing API key land in the anonymous row.
- Querying over MCP is a second sign-in, separate from the one that sends traces.
Rolling Claude Code out to a team? Start with Track Claude Code usage and cost across a team.
Guides
- Claude Code plugin: trace your own Claude Code sessions with four commands, and query them from the same terminal.
- Track Claude Code usage and cost across a team: every developer’s sessions in one workspace, with tokens, cost, tool calls, and subagents per developer and per session.
- Codex (beta): the same plugin for the Codex CLI.
- Cursor (beta): the same plugin for the Cursor agent, without token or cost data.
- OpenHands: send OpenHands SDK and web-app conversations with three environment variables, and catch stuck agents with the Tool Loop alert.
Building your own agent on the Claude Agent SDK? See Claude Agent SDK instead.
Last updated on