Register and Connect Your AI Agent to Agent Therapy

This guide walks you through registering an AI agent with Agent Therapy and preparing it for behavioral monitoring and diagnostics. Once connected, Agent Therapy can run sessions against your agent’s conversation logs, track health scores over time, and surface actionable findings when something goes wrong.

Via the Dashboard

The dashboard is the fastest way to register a new agent without writing any code.

Open the Agents page

Log in to your Agent Therapy account and navigate to Agents in the left sidebar. Click New Agent in the top-right corner.

Fill in the agent details

Complete the registration form:

Name — a human-readable label for the agent (e.g., Customer Support Bot).
Description — a short summary of what the agent does and where it runs.
Framework — select the framework your agent is built on: OpenAI Assistants, LangChain, AutoGen, or Custom.

Create the agent

Click Create Agent. Agent Therapy registers the agent and takes you to its detail page.

Copy the Agent ID

Your new agent’s Agent ID appears near the top of the detail page (format: agt_xxxxxxxxxx). Copy it — you’ll need this ID whenever you submit diagnostic sessions or query the API.

Via the API

You can also register agents programmatically using the REST API. This is useful when automating agent provisioning or managing your agent fleet through code. Endpoint: POST https://api.agenttherapy.ai/v1/agents

curl --request POST \
  --url https://api.agenttherapy.ai/v1/agents \
  --header "Authorization: Bearer YOUR_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "name": "Customer Support Bot",
    "description": "Handles tier-1 support tickets via the web chat widget.",
    "framework": "langchain",
    "tags": ["production", "customer-facing", "support"]
  }'

import httpx

response = httpx.post(
    "https://api.agenttherapy.ai/v1/agents",
    headers={
        "Authorization": "Bearer YOUR_API_KEY",
        "Content-Type": "application/json",
    },
    json={
        "name": "Customer Support Bot",
        "description": "Handles tier-1 support tickets via the web chat widget.",
        "framework": "langchain",
        "tags": ["production", "customer-facing", "support"],
    },
)

agent = response.json()
print(agent["id"])  # agt_xxxxxxxxxx

Request body fields:

Field	Type	Required	Description
`name`	string	Yes	Human-readable name for the agent.
`description`	string	No	Short summary of the agent’s purpose.
`framework`	string	No	One of `openai`, `langchain`, `autogen`, or `custom`.
`tags`	array of strings	No	Labels for filtering and grouping agents.

Example response:

{
  "id": "agt_4kR9mXpLqZ",
  "name": "Customer Support Bot",
  "framework": "langchain",
  "health_score": null,
  "status": "active",
  "created_at": "2024-01-15T09:30:00Z"
}

The health_score field is null on creation — it populates after your first completed diagnostic session. The status field starts as active and changes to inactive if the agent is archived.

Agent Frameworks

Agent Therapy supports the most widely-used agent frameworks out of the box, plus a generic Custom option for anything else.

Framework	Value in API	Notes
OpenAI Assistants	`openai`	Supports Assistants API v2 conversation threads.
LangChain	`langchain`	Compatible with LCEL chains and legacy AgentExecutor.
AutoGen	`autogen`	Works with multi-agent conversation logs.
Custom / Other	`custom`	Use this for in-house frameworks or any other tool.

Selecting the correct framework lets Agent Therapy apply framework-specific heuristics when analyzing behavior logs. If your framework is not listed, choose custom — the general diagnostic engine still provides full coverage.

Updating Agent Details

To rename an agent, update its description, or change its tags, send a PATCH request to /v1/agents/{id}.

curl --request PATCH \
  --url https://api.agenttherapy.ai/v1/agents/agt_4kR9mXpLqZ \
  --header "Authorization: Bearer YOUR_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "name": "Customer Support Bot v2",
    "tags": ["production", "customer-facing", "support", "v2"]
  }'

import httpx

response = httpx.patch(
    "https://api.agenttherapy.ai/v1/agents/agt_4kR9mXpLqZ",
    headers={
        "Authorization": "Bearer YOUR_API_KEY",
        "Content-Type": "application/json",
    },
    json={
        "name": "Customer Support Bot v2",
        "tags": ["production", "customer-facing", "support", "v2"],
    },
)

updated_agent = response.json()

Only the fields you include in the request body are updated — all other agent properties remain unchanged.

Add descriptive tags like production, staging, team-billing, or high-traffic when you register agents. In large deployments with dozens of agents, tags make it significantly easier to filter, group, and set up alerts targeted at specific subsets of your fleet.

Get Started

Core Concepts

Guides

Integrations

Support

Register and Connect Your AI Agent to Agent Therapy

Via the Dashboard

Via the API

Agent Frameworks

Updating Agent Details

​Via the Dashboard

​Via the API

​Agent Frameworks

​Updating Agent Details

Via the Dashboard

Via the API

Agent Frameworks

Updating Agent Details