learning-env

Use relai learning-env to create or upload repeatable scenarios.

`create`

Create a learning environment from a prompt or from a run log plus feedback.

relai learning-env create [OPTIONS] --prompt {PROMPT}
relai learning-env create [OPTIONS] --log-file {PATH} --feedback {TEXT}

Learning environment generation can take a while when the prompt, run log, or agent context is complex.

Option	Description
`--prompt {TEXT}`	Behavior to turn into a learning environment.
`--log-file {PATH}`	Agent run log, in any text format. Requires `--feedback`.
`--feedback {TEXT}`	Feedback about the run. Requires `--log-file`.
`--name {NAME}`	Stable generated file name and environment id.
`--agent-target {TARGET}`	Registered target for this learning environment. Required non-interactively when multiple targets are registered.

relai learning-env create --prompt "{describe the behavior to test}"
relai learning-env create --log-file {path/to/run-log} --feedback "{describe the behavior to preserve}"
relai learning-env create --prompt "{describe the behavior to test}" --name {name}
relai learning-env create --prompt "{describe the behavior to test}" --agent-target {target}

The run log does not need a specific structure: any text format under the size limit works, including JSONL message logs, plain transcripts, or your own trace format. RELAI parses it. Logs that include internal detail — tool calls, retrievals, intermediate reasoning — give RELAI more to work with than user-facing output alone. Creates a learning-environment file under .relai/learning-envs/ using the project’s language. Interactive successful flows may offer to commit and upload it. Non-interactive flows print follow-up commands. Learning environments are scoped to the selected agent target. When multiple targets are registered, interactive runs present a single-target picker; pass --agent-target {target} in non-interactive runs. After upload, open the Agents page, select your agent, and use the Learning Environments tab. Common failures: prompt over 8 KB, feedback over 32 KB, log file over 160 KB, missing paired --feedback, backend generation failure, validation failure, duplicate explicit name, or dirty generated paths.

`upload`

Upload an existing committed learning environment.

relai learning-env upload --learning-envs {name}

relai learning-env upload --learning-envs {name-a},{name-b}
relai learning-env upload --learning-envs {name} --agent-target {target}

Uploads the exact source and git metadata to the RELAI backend. Use --agent-target {target} only to upload a legacy learning environment that does not declare its target in source; a declared target always takes precedence. Common failures: file missing, invalid learning environment, file not tracked or not committed at current HEAD, uncommitted changes, detached HEAD, or backend branch conflict.

Get started

Understand it

Command reference

learning-env

`create`

`upload`

​create

​upload

`create`

`upload`