LoopLlama
Get API access
API reference

Overview

The LoopLlama API is organized around REST. It uses predictable resource URLs, accepts and returns JSON, and relies on standard HTTP verbs and status codes.

Base URL & versioning#

All endpoints live under a single versioned base URL. The current version is v1; breaking changes ship under a new version prefix.

Base URL
https://loopllama.ai/api/v1

Authentication#

Authenticate every request with a Bearer API key. See Authentication for details.

bash
-H "Authorization: Bearer ll_live_..."

Requests#

  • Send request bodies as JSON with a Content-Type: application/json header.
  • Field names in request bodies use camelCase (e.g. systemPrompt), matching the agent schema.
  • Resource ids are passed as path segments, e.g. /v1/workflows/{id}.

Response envelope#

Successful responses wrap the payload in a data key. A single resource returns an object; collections return an array. Response field names use snake_case (e.g. created_at, workflow_id).

json
// Single resource
{ "data": { "id": "ckv9...", "name": "Brief writer" } }

// Collection
{ "data": [ { "id": "ckv9..." }, { "id": "ckva..." } ] }

Errors are returned without a data key — instead they carry a top-level error string. See Errors.

HTTP status codes#

  • 200 OK — a successful read.
  • 201 Created — a workflow was created.
  • 202 Accepted — a run was queued and will execute asynchronously.
  • 400, 401, 404, 409 — client errors, detailed on the Errors page.

Endpoints#

The complete v1 surface:

GET/v1/workflows
POST/v1/workflows
GET/v1/workflows/{id}
GET/v1/workflows/{id}/runs
POST/v1/workflows/{id}/runs
GET/v1/runs/{id}
GET/v1/agents
POST/v1/agents
GET/v1/agents/{id}
PATCH/v1/agents/{id}
DELETE/v1/agents/{id}
GET/v1/agents/{id}/memory
POST/v1/agents/{id}/memory

Resources#

Workflows

List, create, and retrieve workflows and their crews.

Agents

Your standing team: create, pause and edit agents, link them into crews, and read or write their memory.

Runs

Trigger runs, list a workflow's runs, and retrieve a run with its steps.

Webhooks

Register signed HTTP endpoints for run events.

Fine-tunes

Train a base model on your examples or run history, track the job, and route agents to the result.