Get current account status: tier, training credits, prediction usage, endpoint limits. customer_id is optional; the backend resolves the account from the API key.

List all projects for this API key (id, name, description, timestamps). Call this first to obtain project IDs needed by other tools.

Fetch a single project by id.

Create or update a project.

• Create: call with name (and optional description).

• Update/rename: pass project_id of an existing project plus the fields to change (name and/or description).

List datasets in a project (id, name, type, timestamps).

Fetch or update a dataset.

• Read: call with project_id and dataset_id only.

• Update/rename: also pass name and/or description to change.

Datasets are created via complete_upload (the upload flow); use this tool for reads and metadata edits.

Start an upload request and return a gateway upload curl command.

FULL 3-STEP FLOW:

Step 1 — call start_upload. Step 2 — run curl_command in bash. Replace FILE_PATH with the actual file path. Step 3 — call complete_upload with upload_request_id from step 1.

No API key or auth header is needed in the curl command. Pass dataset_id when uploading a new version of an existing dataset.

Finalize an upload after start_upload + curl. upload_request_id: opaque id returned by start_upload. dataset_id: optional target dataset id for creating a new version. If the dataset already exists, a new version is created automatically. dataset_type: train | test | validation (default train).

The gateway PUT from start_upload must return HTTP 2xx before you call this tool; otherwise the API responds with 400 (upload session not UPLOADED yet).

Returns the dataset record with id, name, and the new datasetVersion.

List all versions of a dataset (version number, version_type, qa_status, row counts).

project_id is optional — the backend resolves access from dataset_id when omitted.

Fetch one dataset version by id (metadata, qa_status, version_type).

Create or update a dataset version.

Create (register an S3 file as a new version — used by the QA pipeline): Required: version_type (raw | qa_cleaned | training), file_url (s3:// URL), qa_metadata (freeform JSON with QA results).

Update (change qa_status on an existing version): Required: version_id, qa_status (pending | in_progress | ready | blocked).

Create or update a model.

• Create: call with project_id and name (and optional description).

• Update/rename: also pass model_id plus the fields to change.

Fetch a single model by id (name, description, version count).

Create a model version tied to a dataset version and target column. Then call submit_training_job with the returned model version id.

List all models in a project (id, name).

List model versions. Training state is status (SUBMITTED → TRAINING → TRAINING_COMPLETED or TRAINING_FAILED). Report readiness is edaReportStatus (PENDING | GENERATING | READY | FAILED).

Fetch a single model version by id (status, edaReportStatus, target, timestamps). Prefer this over list_model_versions when you already know the version_id.

Load the EDA training report (metrics, feature analysis, performance summary).

project_id is optional — the backend resolves it from model_id when omitted.

Omit model_version_id to use the latest version for this model. Pass an id from list_model_versions to read a specific version.

Default response is summary only (token-efficient). Set full_report=true for full detail.

Submit a training job for a model version.

Track completion: Poll list_model_versions for the model and watch the target version's status until it reaches TRAINING_COMPLETED or TRAINING_FAILED (same as the web UI). This is the most reliable approach across MCP hosts.

If your integration exposes get_training_status, you can instead pass the returned jobId with wait=true to block until the Batch job finishes (typical 2–3 min). Some hosts omit that tool when running stale server code — use list_model_versions then.

dataset_version_id can be omitted when the model version was created with create_model_version in the same flow — the backend resolves target_feature, file, and dataset from the model version record automatically. Pass it only to override with a different qa_cleaned/training dataset version.

Returns {jobId, modelVersionId, status}.

Check a training job by job_id (the jobId field from submit_training_job).

If this tool does not appear in your MCP tool list: restart the host and ensure the client runs current easydeploy_ai_mcp (standard catalog is 24 tools). Until then, poll list_model_versions for the model version status instead.

Response fields:

• status: PENDING | RUNNING | COMPLETE | FAILED

• trainingTimeSeconds: wall-clock seconds once the job stops; null while running

• modelVersionId: the model version being trained

By default returns the current status immediately.

Set wait=true to block until the job reaches a terminal state (COMPLETE or FAILED). Polls every poll_interval_seconds (default 10 s) for up to timeout_seconds (default 180 s / 3 min). Typical training runs finish in 2-3 minutes. If the timeout expires, the last polled status is returned with timed_out: true.

Run a single ad-hoc prediction against a trained model version.

project_id and target_feature are auto-resolved from the model version record when omitted. By default waits and returns the result inline (label + probability). Set wait_for_result=false to return immediately with prediction_id.

Score an entire dataset against a trained model version.

project_id and target_feature are auto-resolved from the model version record when omitted. dataset_version_id identifies both the input file and the row count for credit billing.

Returns immediately by default (fire-and-poll). Use get_prediction(prediction_id) to check status (includes downloadReady flag for completed batches). Set wait_for_result=true to block.

Fetch prediction status and result by prediction id.

• Ad-hoc completed: output contains the label/probability.

• Batch completed: download_url and curl_command are included automatically (tokenized gateway proxy, safe from remote sandbox).

List predictions (newest first). Optionally filter by project_id. Use get_prediction(id) for full result + batch download URL.