Rudel

Rudel's core idea is simple: your coding agents already write down everything they do, so why not read it back properly? It ingests Claude Code and Codex session transcripts and turns each one into a browsable timeline. You can follow prompts, responses, tool calls, commands, and subagent activity in the order they occurred, which makes it far easier to see exactly where and why a long agent run went sideways instead of scrolling through raw log files by hand.

The product's original hook was a playful one, framed around "trading cards" for your coding sessions. Rudel keeps that spirit with recaps that summarize how you've used your agents, which model did the heavy lifting, and where the effort went. It's a lighter way to look at your own habits before you dig into the full dashboards. Fun first, useful second.

Rudel is built around a shared workspace, not a single developer's terminal. Once sessions sync, the team can browse activity across repositories and see coding-agent adoption, activity, token volume, and estimated spend in one place. For leads, that means a real answer. Is the team actually using these tools, and what is it costing us? A guess won't survive a budget meeting.

Every session opens into a full trace: prompts, responses, tool calls, commands, and subagent activity in sequence. That level of detail matters. When something goes wrong and you need to reconstruct the reasoning, a summary won't cut it. Rather than reading a summary, you walk the actual path the agent took, which is usually where you find the prompt or file that caused the problem.

The CLI organizer groups sessions by repository, including work from Git worktrees, and shows local versus uploaded counts side by side so you can see at a glance what has already made it into the workspace and what still lives only on someone's laptop. You decide which repositories sync. Sessions already uploaded are skipped, so re-running setup doesn't create duplicates. It's a deliberately narrow control surface, and that's the point.

On the paid tiers, Rudel surfaces recurring patterns, blockers, and failures across an organization's sessions. Instead of reading sessions one at a time, you look for the friction that keeps showing up. The weekly digest is the lighter version of the same idea, a regular report on adoption, performance changes, and notable sessions.

The Enterprise tier adds on-prem or self-hosted deployment, invoice and PO billing, SAML and SCIM, granular admin controls, and migration support. That's the tier aimed at organizations that can't send session data to a third-party cloud. For everyone else, the cloud workspace is the default path.