Skip to content
Docs
Esc
↑↓navigate↵open⌘Jpreview
On this page

Console

This page covers deepgym console, the local web console for the environments, runs, episodes, sandboxes and QA reports on your machine: what each page shows, where data is stored, and how the console protects itself.

pip install "deepgym[server]"
deepgym console --port 8080 --allow-local-exec   # http://127.0.0.1:8080

deepgym web is an alias.

Pages

  • Environments: search and filter the registry; open one to read its task, verifier, reference solution and metadata.

  • Runs: score a pasted solution or the reference against one environment or a small eval, on any provider (local, docker, apple, daytona, modal, e2b). Runs execute in the background; the run page shows per-case results and a diff against the reference. A run without a verdict keeps no score and counts as an error, and the totals show a “No verdict” count with the judged pass rate and score beside the headline ones.

  • Episodes: replay recorded trajectories step by step, with screenshots for computer-use tasks. Record one from your agent loop:

    from deepgym.console import record_episode

trajectory, result = dg.run_episode(env, agent)
record_episode(trajectory, result, environment='todo_list')

    record_episode(tr.trace) stores a trace the same way, and keeps it for deepgym traces export.

  • Sandboxes: live deepgym-* containers from Docker and Apple container, with kill. Cloud sandboxes are not listed.

  • QA: the reports deepgym qa writes to ~/.deepgym/qa/ (or $DEEPGYM_QA_DIR).

Storage and execution

Runs and episodes are stored in ~/.deepgym/console.db. Reference solutions may run locally, but a pasted solution only runs on your machine with --allow-local-exec (or DEEPGYM_ALLOW_LOCAL_EXEC=true); otherwise pick a sandbox provider.

Security

The console has no login, so other websites can’t drive it:

  • Every /api call needs the per-launch token the console writes into its pages, sent as X-DeepGym-Token.
  • Writes from other origins are refused, and so are write bodies that aren’t JSON.
  • Bound to 127.0.0.1 or localhost (the default), it only answers requests addressed to localhost. Bound to any other host, it doesn’t check the Host header.

Building the UI

The UI lives in console-ui/ (Next.js static export, built with bun) and ships prebuilt in wheels. From a clone, build it once with cd console-ui && bun install && bun run build.