beam-ai-team/get-beam-agent-graph

Get Beam Agent Graph

Fetch agent graph JSON from Beam.ai API and save to local storage with meaningful filename.

Purpose

This skill retrieves the complete agent graph structure from Beam.ai including nodes, edges, configurations, and metadata. It authenticates with the Beam API, fetches the graph for a specific agent, and stores it locally for analysis, backup, or documentation purposes.

Key Features:

• Two-step authentication (API key → access token → graph data)
• Automatic meaningful filename generation (includes agent name if available)
• Organized storage in 04-workspace/beam-graphs/
• Full graph structure preserved in JSON format

Time Estimate: 1-2 minutes

---

Workflow

Step 1: Verify Configuration

Check if BEAM_API_KEY exists in .env:

grep "BEAM_API_KEY" ../.env

If missing:

[ERROR] BEAM_API_KEY not found in .env

Please add your Beam API key to .env:
BEAM_API_KEY=your-key-here

Ask user to provide their Beam API key and add it to .env file.

---

Step 2: Get Required Parameters

Required inputs:

• Workspace ID (UUID format)
• Agent ID (UUID format)

Ask user if not provided:

To fetch the agent graph, I need:
1. Workspace ID: <uuid>
2. Agent ID: <uuid>

Please provide both IDs.

---

Step 3: Execute Fetch Script

Run the get_agent_graph.py script:

python3 03-skills/get-beam-agent-graph/scripts/get_agent_graph.py <workspace_id> <agent_id>

Optional: Specify custom output directory:

python3 03-skills/get-beam-agent-graph/scripts/get_agent_graph.py <workspace_id> <agent_id> --output ./custom/path

The script will:

1. Load BEAM_API_KEY from .env
2. Call /auth/access-token to get access token
3. Call /agent-graphs/{agentId} to fetch graph
4. Save to 04-workspace/beam-graphs/{filename}.json

---

Step 4: Verify and Report Success

On success:

[SUCCESS] Agent graph saved to: 04-workspace/beam-graphs/AgentName_162e7c30_graph_20260105_143022.json

Read the saved file to confirm:

cat 04-workspace/beam-graphs/<filename>.json | head -20

Show user:

• File path
• Agent name (if extracted from graph)
• File size
• Brief summary of graph structure (node count, etc.)

---

API Reference

For detailed API documentation, see references/api-reference.md.

Quick reference:

Step 1: Get Access Token

• Endpoint: POST https://api.beamstudio.ai/auth/access-token
• Header: x-api-key: <BEAM_API_KEY>
• Returns: access_token

Step 2: Get Agent Graph

• Endpoint: GET https://api.beamstudio.ai/agent-graphs/{agentId}
• Headers: Authorization: Bearer <token>, x-workspace-id: <workspace_id>
• Returns: Complete agent graph JSON

---

Output Format

Filename pattern:

• With agent name: {AgentName}_{agentId_first8}graph_{timestamp}.json
• Without name: {agentId}_graph_{timestamp}.json

Example filenames:

TrueSearch_Scheduler_162e7c30_graph_20260105_143022.json
162e7c30-0d95-49ab-af99-7eef872a2d0d_graph_20260105_143022.json

Storage location:

04-workspace/beam-graphs/

---

Error Handling

Common errors:

| Error | Cause | Solution | |-------|-------|----------| | BEAM_API_KEY not found | Missing from .env | Add to .env file | | 401 Unauthorized | Invalid API key | Check BEAM_API_KEY value | | 404 Not Found | Invalid agent ID | Verify agent ID is correct | | 403 Forbidden | No workspace access | Check workspace ID and permissions |

For API details: See references/api-reference.md

---

Resources

scripts/

scripts/get_agent_graph.py - Main script to fetch and save agent graphs

• Handles authentication flow
• Fetches graph via Beam API
• Saves with meaningful filename
• Supports custom output directory

references/

references/api-reference.md - Beam API documentation

• Authentication endpoints
• Agent graph endpoints
• Error codes and responses

---

Notes

About Beam API Authentication:

• Access tokens are short-lived (typically 1 hour)
• The script requests a new token for each execution
• API key is stored in .env and never logged

About Storage:

• Graphs are timestamped to prevent overwrites
• Default location: 04-workspace/beam-graphs/
• Custom output directory supported via --output flag

About Graph Data:

• Complete graph structure including nodes, edges, configs
• JSON format for easy parsing and analysis
• Preserves all metadata from Beam API