lerobot-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| LEROBOT_ROOT | No | Optional explicit path to a LeRobot checkout. If not set, the server searches in standard locations: current project ancestors, managed checkout ~/.cache/lerobot-mcp/lerobot, ~/hrl/lerobot, and installed lerobot package. | |
| FORGE_REGISTRY_PATH | No | Optional path to a local datasets.json registry file for offline or deterministic dataset search. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| lerobot_server_configA | Return resolved server configuration and LeRobot checkout paths. |
| lerobot_install_or_update_lerobotB | Clone or fast-forward update a LeRobot checkout, then refresh this MCP server's config. |
| lerobot_list_commandsB | List supported command names and discovered LeRobot console scripts. |
| lerobot_capabilitiesA | List LeRobot scripts, extras, examples, and registered components from the current checkout. |
| lerobot_public_symbolsA | List public classes/functions below a LeRobot module prefix by static source inspection. |
| lerobot_list_examplesC | List Python example scripts under the configured LeRobot checkout. |
| lerobot_build_commandB | Preview a known LeRobot entry point command without running it. |
| lerobot_command_helpB | Run |
| lerobot_build_exampleB | Preview a LeRobot example script command without running it. |
| lerobot_run_commandC | Run a known LeRobot entry point as a foreground call or managed background job. |
| lerobot_run_exampleC | Run a Python script from LeRobot's examples directory. |
| lerobot_build_forge_convertC | Preview pinned Forge dataset conversion to or from LeRobot-compatible formats. |
| lerobot_forge_convertC | Run pinned Forge dataset conversion. Uses Forge main at the pinned bug-fix commit. |
| lerobot_forge_inspectC | Inspect a dataset with pinned Forge, useful before conversion. |
| lerobot_build_forge_inspectD | Preview pinned Forge dataset inspection. |
| lerobot_build_dataset_latest_format_convertB | Preview LeRobot's official v2.1 -> current v3.0 parquet dataset conversion command. |
| lerobot_convert_dataset_to_latest_formatA | Convert a LeRobot v2.1 dataset to the current v3.0 parquet layout. Defaults to local/no-push behavior. Set |
| lerobot_job_statusA | Return state, return code, and output tails for a background job. |
| lerobot_list_jobsA | List background jobs started during this MCP server process. |
| lerobot_job_logsC | Return stdout and stderr tails for a background job. |
| lerobot_cancel_jobC | Terminate a background job. |
| lerobot_inspect_dataset_metadataC | Summarize LeRobot dataset metadata from a local path or Hugging Face dataset repo. |
| lerobot_hf_whoamiA | Return the Hugging Face account visible to huggingface_hub. |
| lerobot_hf_repo_infoC | Return basic Hugging Face Hub repository information. |
| lerobot_hf_search_datasetsB | Search robotics datasets by robot, format, scale, size, task, and compatibility hints. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 25 tools
Each tool has a distinct purpose, with clear differentiation between preview (build_*) and execution (run_*) commands, as well as separate utilities for Hugging Face, job management, and configuration. Overlapping verbs like 'build' and 'run' are disambiguated by their noun suffixes.
All tools share the 'lerobot_' prefix, and most follow a 'verb_noun' pattern (e.g., build_command, run_example). Minor deviations exist (lerobot_capabilities, lerobot_public_symbols) but the overall pattern is consistent and predictable.
25 tools cover a wide range of LeRobot functionalities—installation, execution, dataset conversion, Hugging Face interaction, job management—without feeling excessive. The count is appropriate for the server's scope, though slightly higher than ideal.
The tool surface addresses core LeRobot workflows: running commands and examples, converting datasets, inspecting metadata, and managing background jobs. Missing upload or training tools are likely outside the intended scope, so the set feels reasonably complete.