Skip to content

Loom API Capabilities Summary

Date: 2026-01-22
System Status: Running and Healthy
Authentication: admin/admin (working, token-based JWT)

Executive Summary

Yes to all your questions!

  1. ✅ All UI features are implemented through API calls
  2. ✅ You can call the CEO/agents via API/curl
  3. ✅ You can invoke agents via API to review the codebase
  4. ✅ You can check provider status via API

Current System Status

Running Containers

- loom (port 8080) - HEALTHY

Active Agents (Currently 10/10 - at capacity)

1. agent-1769059130-Project Manager (Default) - paused
2. agent-1769059130-Documentation Manager (Default) - paused
3. agent-1769059130-Web Designer (Default) - paused ⭐
4. agent-1769059130-CFO (Default) - paused
5. agent-1769059130-QA Engineer (Default) - paused
6. agent-1769059130-Devops Engineer (Default) - paused
7. agent-1769059130-Code Reviewer (Default) - paused
8. agent-1769059130-CEO (Default) - paused ⭐
9. agent-1769059130-Product Manager (Default) - paused
10. agent-1769059130-Engineering Manager (Default) - paused

Note: All agents are currently PAUSED because the only registered provider is a mock provider (mock-local) which doesn't support real work.

Registered Providers

{
  "id": "mock-local",
  "name": "Local Mock Provider",
  "type": "mock",
  "status": "active",
  "endpoint": "mock://local"
}

Action needed: Register a real LLM provider to enable agent work.

API Endpoints - Complete Coverage

Authentication & Authorization ✅

# Login
POST /api/v1/auth/login
{"username":"admin","password":"admin"}

# Get current user
GET /api/v1/auth/me

# Change password
POST /api/v1/auth/change-password

# Generate API key
POST /api/v1/auth/api-keys

# List users (admin)
GET /api/v1/auth/users

Provider Management ✅

# List providers (includes health status)
GET /api/v1/providers
Authorization: Bearer <token>

# Register new provider
POST /api/v1/providers
{
  "name": "My Provider",
  "type": "ollama",
  "endpoint": "http://localhost:11434",
  "model": "llama3.2",
  "api_key": "optional"
}

# Get provider details
GET /api/v1/providers/{id}

# Get provider models
GET /api/v1/providers/{id}/models

# Delete provider
DELETE /api/v1/providers/{id}

Agent Management ✅

# List all agents
GET /api/v1/agents

# Spawn new agent
POST /api/v1/agents
{
  "persona_name": "default/web-designer",
  "project_id": "loom-self",
  "provider_id": "mock-local"
}

# Get agent details
GET /api/v1/agents/{id}

# Stop agent
DELETE /api/v1/agents/{id}

# Pause/resume agent
POST /api/v1/agents/{id}/pause
POST /api/v1/agents/{id}/resume

CEO REPL (Direct Agent Invocation) ✅

# Ask the CEO agent a question
POST /api/v1/repl
{
  "message": "Review the web UI and provide recommendations",
  "timeout_sec": 180
}

# Returns a response from the CEO persona

Work Submission (Non-Bead Prompts) ✅

# Submit ad-hoc work to an agent
POST /api/v1/work
{
  "prompt": "Analyze the current web UI design",
  "agent_id": "optional-specific-agent",
  "project_id": "loom-self"
}

Personas ✅

# List all personas
GET /api/v1/personas

# Get specific persona
GET /api/v1/personas/default/web-designer

# Update persona (live editing)
PUT /api/v1/personas/{name}

Projects ✅

# List projects
GET /api/v1/projects

# Get project details
GET /api/v1/projects/loom-self

# Create project
POST /api/v1/projects

# Sync git
POST /api/v1/projects/git/sync
POST /api/v1/projects/git/commit
POST /api/v1/projects/git/push

Beads (Work Items) ✅

# List beads
GET /api/v1/beads?project_id=loom-self&status=open

# Create bead
POST /api/v1/beads

# Update bead
PATCH /api/v1/beads/{id}

# Claim bead (assign to agent)
POST /api/v1/beads/{id}/claim

Decisions ✅

# List decision beads
GET /api/v1/decisions

# Make a decision
POST /api/v1/decisions/{id}/decide

System Status ✅

# Get overall system status
GET /api/v1/system/status

# Health check
GET /api/v1/health

Analytics ✅

# Get usage logs
GET /api/v1/analytics/logs

# Get statistics
GET /api/v1/analytics/stats

# Get cost report
GET /api/v1/analytics/costs

# Export data
GET /api/v1/analytics/export

Web UI Implementation

The web UI is a single-page application that uses 100% API calls for all functionality:

UI Sections (All API-Backed)

  1. Project Viewer - /api/v1/projects, /api/v1/beads, /api/v1/agents
  2. Kanban Board - /api/v1/beads with filtering
  3. Providers - /api/v1/providers (registration, health checks)
  4. Agents - /api/v1/agents (spawning, monitoring)
  5. Personas - /api/v1/personas (viewing, live editing)
  6. Projects - /api/v1/projects (CRUD operations)
  7. Decisions - /api/v1/decisions (decision making)
  8. CEO REPL - /api/v1/repl (direct agent queries) ⭐
  9. Streaming Test - /api/v1/chat/completions/stream
  10. Users - /api/v1/auth/users (user management)
  11. Analytics - /api/v1/analytics/* (dashboards, reports)

Key Features

  • Real-time updates via Server-Sent Events (EventSource)
  • Streaming responses for chat completions
  • Auto-refresh every 5 seconds
  • JWT authentication with token refresh
  • Modal dialogs for forms
  • Toast notifications for feedback

How to Invoke Agents via curl

1. Check Provider Status First

# Login
TOKEN=$(curl -s -X POST http://localhost:8080/api/v1/auth/login \
  -H "Content-Type: application/json" \
  -d '{"username":"admin","password":"admin"}' | jq -r '.token')

# Check providers
curl -s -H "Authorization: Bearer $TOKEN" \
  http://localhost:8080/api/v1/providers | jq .

2. Register a Real Provider (if needed)

# Example: Register Ollama
curl -X POST http://localhost:8080/api/v1/providers \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Ollama Local",
    "type": "ollama",
    "endpoint": "http://localhost:11434",
    "model": "llama3.2"
  }' | jq .

3. Invoke CEO Agent via REPL

# Ask CEO to review the codebase
curl -X POST http://localhost:8080/api/v1/repl \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Please review the Loom web UI codebase and provide recommendations for improvements. Focus on accessibility, usability, and modern web standards.",
    "timeout_sec": 300
  }' | jq .

4. Ask Web Designer to Review

Option A: Via existing agent (currently paused)

First, we need to resume or create a new web designer agent. Since we're at capacity, let's use the REPL to route to the appropriate persona:

# The CEO can delegate to web designer internally
curl -X POST http://localhost:8080/api/v1/repl \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Delegate to the web-designer persona: Please review our web UI at /web/static/ and provide a detailed UX/UI assessment with specific recommendations.",
    "timeout_sec": 300
  }' | jq .

Option B: Remove an agent and spawn web designer

# Stop an agent to free up capacity
curl -X DELETE http://localhost:8080/api/v1/agents/agent-1769059130-CFO%20(Default) \
  -H "Authorization: Bearer $TOKEN"

# Spawn web designer with a real provider (once registered)
curl -X POST http://localhost:8080/api/v1/agents \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "persona_name": "default/web-designer",
    "project_id": "loom-self",
    "provider_id": "your-real-provider-id"
  }' | jq .

5. Submit Direct Work

# Submit work directly without creating a bead
curl -X POST http://localhost:8080/api/v1/work \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Review web/static/index.html and web/static/js/app.js. Evaluate accessibility, performance, and UX. Provide 5 specific, actionable recommendations.",
    "project_id": "loom-self"
  }' | jq .

Available Personas

Executive/Management

  • default/ceo - Strategic decisions, high-level planning
  • default/cfo - Financial analysis, budgeting
  • default/product-manager - Product strategy, roadmaps
  • default/project-manager - Task coordination, timelines
  • default/engineering-manager - Team coordination, technical leadership
  • default/decision-maker - Decision resolution

Engineering

  • default/code-reviewer - Code quality, best practices
  • default/devops-engineer - Infrastructure, deployment
  • default/qa-engineer - Testing, quality assurance

Design & Content

  • default/web-designer ⭐ - UI/UX design, accessibility
  • default/web-designer-engineer ⭐ - Full-stack design+implementation
  • default/documentation-manager - Documentation quality
  • default/public-relations-manager - Communication, messaging

Operations

  • default/housekeeping-bot - Maintenance, cleanup

Current Limitations

1. Provider Required

  • Mock provider doesn't execute real work
  • Need to register: Ollama, OpenAI, or other LLM provider
  • Health checks run automatically on registration

2. Agent Capacity

  • Currently at 10/10 agents (maximum)
  • Need to stop agents to spawn new ones
  • OR increase agents.max_concurrent in config.yaml

3. Agents Paused

  • All agents paused due to no viable provider
  • Will auto-resume when real provider is active

Immediate: Check Provider Health

TOKEN=$(curl -s -X POST http://localhost:8080/api/v1/auth/login \
  -H "Content-Type: application/json" \
  -d '{"username":"admin","password":"admin"}' | jq -r '.token')

curl -s -H "Authorization: Bearer $TOKEN" \
  http://localhost:8080/api/v1/providers | \
  jq '.[] | {id, name, status, last_heartbeat_at, last_heartbeat_error}'

Next: Register Real Provider

# If you have Ollama running locally
curl -X POST http://localhost:8080/api/v1/providers \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "id": "ollama-local",
    "name": "Ollama Local LLM",
    "type": "ollama",
    "endpoint": "http://host.docker.internal:11434",
    "model": "llama3.2",
    "description": "Local Ollama instance"
  }' | jq .

Then: Invoke Web Designer Review

# Method 1: Via CEO REPL (delegates internally)
curl -X POST http://localhost:8080/api/v1/repl \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Have the web-designer persona conduct a comprehensive review of our web UI (web/static/). Focus on: 1) Accessibility compliance, 2) Mobile responsiveness, 3) Visual hierarchy, 4) User interaction patterns, 5) Performance considerations. Provide specific, actionable recommendations.",
    "timeout_sec": 300
  }' | jq .

# Method 2: Direct work submission
curl -X POST http://localhost:8080/api/v1/work \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "As a web designer, review the Loom UI and provide a detailed UX assessment.",
    "project_id": "loom-self",
    "persona_hint": "web-designer"
  }' | jq .

Summary

Everything you asked about is possible!

  1. ✅ UI is 100% API-driven - Every feature uses REST endpoints
  2. ✅ CEO/CLI callable via curl - /api/v1/repl endpoint
  3. ✅ Web designer can review codebase - Via REPL or direct work submission
  4. ✅ Provider health check available - /api/v1/providers shows status

Blockers: - Need to register a real LLM provider (mock provider insufficient) - Agents currently paused waiting for viable provider

Next Steps: 1. Register an LLM provider (Ollama, OpenAI, etc.) 2. Verify provider status shows "active" 3. Invoke web-designer via API 4. Review results

Example Complete Workflow

#!/bin/bash
# Complete workflow to review web UI via API

# 1. Login
TOKEN=$(curl -s -X POST http://localhost:8080/api/v1/auth/login \
  -H "Content-Type: application/json" \
  -d '{"username":"admin","password":"admin"}' | jq -r '.token')

echo "Logged in: $TOKEN"

# 2. Check providers
echo "\n=== Checking providers ==="
curl -s -H "Authorization: Bearer $TOKEN" \
  http://localhost:8080/api/v1/providers | jq '.[] | {id, name, status}'

# 3. Check agents
echo "\n=== Checking agents ==="
curl -s -H "Authorization: Bearer $TOKEN" \
  http://localhost:8080/api/v1/agents | jq '.[] | {id, persona_name, status}'

# 4. If provider is active, ask web designer to review
echo "\n=== Requesting web UI review ==="
curl -X POST http://localhost:8080/api/v1/repl \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Web designer: Please review web/static/index.html, web/static/css/style.css, and web/static/js/app.js. Provide 5 specific UX improvements.",
    "timeout_sec": 300
  }' | jq .

Generated: 2026-01-22 at 13:28 UTC
System: Loom v1.0.0
API Version: v1
Authentication: JWT Bearer Token