Gradient Labs

Gradient Labs builds "Otto", an AI customer support agent that autonomously handles complex, end-to-end support conversations for regulated and financial services businesses. The API lets you start and drive conversations, stream customer messages, hand off to human agents, execute business actions/tools, and manage the knowledge base the agent reasons over, with signed webhooks delivering the agent's outbound messages and events.

6 APIs 0 Features

AICustomer SupportAI AgentConversationsFinancial ServicesRegulated

APIs

Gradient Labs Conversations API

Start a conversation, assign it to the Gradient Labs AI agent, read its current state, and manage its lifecycle (finish, cancel, resume). The AI agent autonomously works the con...

Gradient Labs Messages API

Add inbound customer and human-agent messages (with optional email subject and attachments) to an existing conversation for the AI agent to process.

Gradient Labs Hand-off & Assignment API

Assign or reassign a conversation between the AI agent and human participants, and capture the conversation.hand_off webhook the agent emits when it decides to escalate to a hum...

Gradient Labs Actions & Tools API

Define the business tools the AI agent can call and execute them. Tools can be run via the API or delivered to your endpoint through the action.execute webhook so the agent can ...

Gradient Labs Knowledge API

Upsert and delete knowledge-base articles (title, description, body, visibility, topic, status) that ground the AI agent's answers, organized by topic.

Gradient Labs Webhooks API

Signed, ordered webhook events the AI agent delivers to your endpoint - agent.message, conversation.hand_off, conversation.finished, action.execute, and resource.pull - each car...

Collections

Gradient Labs API

OPEN

Pricing Plans

Gradient Labs Plans Pricing

1 plans

PLANS

Rate Limits

Gradient Labs Rate Limits

2 limits

RATE LIMITS

FinOps

Gradient Labs Finops

FINOPS

Resources

GitHubOrganization

Sources

opencollection: 1.0.0
info:
  name: Gradient Labs API
  version: '1.0'
request:
  auth:
    type: bearer
    token: '{{apiKey}}'
items:
- info:
    name: Conversations
    type: folder
  items:
  - info:
      name: Start a conversation
      type: http
    http:
      method: POST
      url: https://api.gradient-labs.ai/conversations
      body:
        type: json
        data: "{\n  \"id\": \"conversation-1234\",\n  \"customer_id\": \"customer-5678\",\n  \"channel\": \"chat\",\n  \"\
          assignee_type\": \"ai-agent\"\n}"
    docs: Creates a new conversation and optionally assigns it to the Gradient Labs AI agent.
  - info:
      name: Read a conversation
      type: http
    http:
      method: GET
      url: https://api.gradient-labs.ai/conversations/:conversationID/read
      params:
      - name: conversationID
        value: ''
        type: path
        description: Your unique identifier for the conversation.
    docs: Returns the current state of a conversation.
  - info:
      name: Finish a conversation
      type: http
    http:
      method: PUT
      url: https://api.gradient-labs.ai/conversations/:conversationID/finish
      params:
      - name: conversationID
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Marks a conversation as finished.
  - info:
      name: Cancel a conversation
      type: http
    http:
      method: PUT
      url: https://api.gradient-labs.ai/conversations/:conversationID/cancel
      params:
      - name: conversationID
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Cancels a conversation, stopping the AI agent from working it.
  - info:
      name: Resume a conversation
      type: http
    http:
      method: PUT
      url: https://api.gradient-labs.ai/conversations/:conversationID/resume
      params:
      - name: conversationID
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Resumes a previously finished or cancelled conversation.
  - info:
      name: Rate a conversation
      type: http
    http:
      method: PUT
      url: https://api.gradient-labs.ai/conversations/:conversationID/rate
      params:
      - name: conversationID
        value: ''
        type: path
      body:
        type: json
        data: "{\n  \"type\": \"csat\",\n  \"value\": 5,\n  \"min_value\": 1,\n  \"max_value\": 5\n}"
    docs: Records a customer satisfaction rating for a conversation.
- info:
    name: Messages
    type: folder
  items:
  - info:
      name: Add a message to a conversation
      type: http
    http:
      method: POST
      url: https://api.gradient-labs.ai/conversations/:conversationID/messages
      params:
      - name: conversationID
        value: ''
        type: path
      body:
        type: json
        data: "{\n  \"id\": \"message-1\",\n  \"body\": \"Hi, I need help with a payment dispute.\",\n  \"participant_id\"\
          : \"customer-5678\",\n  \"participant_type\": \"customer\"\n}"
    docs: Adds an inbound customer or human-agent message for the AI agent to process.
- info:
    name: Hand-off
    type: folder
  items:
  - info:
      name: Assign a conversation
      type: http
    http:
      method: PUT
      url: https://api.gradient-labs.ai/conversations/:conversationID/assignee
      params:
      - name: conversationID
        value: ''
        type: path
      body:
        type: json
        data: "{\n  \"assignee_type\": \"ai-agent\",\n  \"reason\": \"Routing to the AI agent\"\n}"
    docs: Assigns or reassigns a conversation between the AI agent and human participants.
- info:
    name: Actions & Tools
    type: folder
  items:
  - info:
      name: Create a tool
      type: http
    http:
      method: POST
      url: https://api.gradient-labs.ai/tools
      body:
        type: json
        data: "{\n  \"id\": \"lookup-account\",\n  \"name\": \"Lookup Account\",\n  \"description\": \"Fetch an account by\
          \ customer id.\"\n}"
    docs: Defines a business tool the AI agent can call. Requires a Management API key.
  - info:
      name: Execute a tool
      type: http
    http:
      method: POST
      url: https://api.gradient-labs.ai/tools/:id/execute
      params:
      - name: id
        value: ''
        type: path
        description: The tool identifier.
      body:
        type: json
        data: "{\n  \"id\": \"lookup-account\",\n  \"arguments\": [\n    { \"name\": \"customer_id\", \"value\": \"customer-5678\"\
          \ }\n  ]\n}"
    docs: Executes a defined tool and returns its JSON-encoded result. Requires a Management API key.
- info:
    name: Knowledge
    type: folder
  items:
  - info:
      name: Upsert a knowledge article
      type: http
    http:
      method: POST
      url: https://api.gradient-labs.ai/articles
      body:
        type: json
        data: "{\n  \"id\": \"refunds-policy\",\n  \"title\": \"Refunds Policy\",\n  \"description\": \"How refunds are handled.\"\
          ,\n  \"body\": \"...\",\n  \"visibility\": \"public\",\n  \"status\": \"published\"\n}"
    docs: Creates or updates a knowledge-base article that grounds the AI agent.
bundled: true