Toolhouse

aid: toolhouse
url: https://raw.githubusercontent.com/api-evangelist/toolhouse/refs/heads/main/apis.yml
apis:
- aid: toolhouse:platform-api
  name: Toolhouse Platform API
  tags:
  - Agent Runs
  - AI Agents
  - AI Workers
  - Agent Studio
  - API Keys
  - Billing
  - Credits
  - Integrations
  - Management
  - Metrics
  - Monetization
  - Schedules
  - SDK
  humanURL: https://docs.toolhouse.ai/toolhouse
  properties:
  - url: https://docs.toolhouse.ai/toolhouse
    type: Documentation
  - url: openapi/toolhouse-openapi-original.yml
    type: OpenAPI
  - url: rules/toolhouse-rules.yml
    type: SpectralRules
  description: The Toolhouse Platform API provides management and orchestration capabilities for AI workers (also referred
    to as AI agents) and their connected tools. It exposes endpoints for user profile and billing administration, API key
    lifecycle, worker/agent definition and deployment, agent-run management with paginated listing and per-run logs (including
    MCP server logs), Agent Studio chat sessions, agent file management, agent subscriptions and monetization, agent transfer,
    scheduled execution via cron expressions (10-minute minimum interval), MCP server and MCP registry configuration, integrations
    and auth-connection management, OAuth callback handling, Stripe webhook and session handling, and an agent-runs metrics
    surface for usage, volume, and per-agent summary reporting. The SDK API exposes the same agent and schedule primitives
    under /v1/ for direct programmatic use by Toolhouse SDK clients. Authentication uses HTTPBearer JWT tokens and per-user
    API keys.
- aid: toolhouse:workers-api
  name: Toolhouse Workers API
  tags:
  - AI Agents
  - AI Workers
  - Attachments
  - Conversations
  - Execution
  - Statefulness
  - Streaming
  humanURL: https://docs.toolhouse.ai/toolhouse/developers/workers-api
  baseURL: https://agents.toolhouse.ai
  properties:
  - url: https://docs.toolhouse.ai/toolhouse/developers/workers-api
    type: Documentation
  - url: https://docs.toolhouse.ai/toolhouse/developers/authentication
    type: Authentication
  - url: https://docs.toolhouse.ai/toolhouse/developers/worker-statefulness
    type: Documentation
  - url: https://docs.toolhouse.ai/toolhouse/developers/attachments
    type: Documentation
  description: The Toolhouse Workers API (historically the "Agents API") enables HTTP execution of any deployed Toolhouse
    worker. Workers defined via the Agent Editor or as code are reachable at https://agents.toolhouse.ai/{agent_id} with POST
    to start a new task, PUT /{agent_id}/{run_id} to continue an existing conversation, and GET /{agent_id}/{run_id} to retrieve
    full history. NDJSON variants under /ndjson/ stream the worker's output plus out-of-band tool-call and debug signals.
    Conversation continuity is keyed by the X-Toolhouse-Run-ID response header. Attachments are supported via URL (30-second
    download timeout) or base64-encoded direct upload (10 MB max per attachment). Public workers require no authentication;
    private workers use HTTPBearer with a Toolhouse API key.
name: Toolhouse
tags:
- Agent Infrastructure
- AI Agents
- AI Workers
- Backend as a Service
- MCP
- MCP Discovery
- RAG
- Tools
- Workers API
type: Index
image: https://kinlane-images.s3.amazonaws.com/shared/apis-json/apis-json-logo.jpg
access: 3rd-Party
common:
- url: https://toolhouse.ai/
  name: Toolhouse Website
  type: Website
  description: 'null'
- url: https://docs.toolhouse.ai/toolhouse
  name: Toolhouse Documentation
  type: Documentation
  description: 'null'
- url: https://toolhouse.ai/blog
  name: Toolhouse Blog
  type: Blog
  description: 'null'
- url: https://toolhouse.ai/pricing
  name: Toolhouse Pricing
  type: Pricing
  description: 'null'
- url: https://app.toolhouse.ai
  name: Toolhouse Login
  type: Login
  description: 'null'
- url: https://toolhouse.ai/about
  name: Toolhouse About
  type: About
  description: 'null'
- url: https://github.com/toolhouseai
  name: Toolhouse GitHub
  type: GitHubOrganization
  description: 'null'
- url: https://toolhouse.ai/privacy
  name: Toolhouse Privacy Policy
  type: PrivacyPolicy
  description: 'null'
- url: https://toolhouse.ai/tos
  name: Toolhouse Terms of Service
  type: TermsOfService
  description: 'null'
- url: https://x.com/toolhouseai
  name: Toolhouse Twitter
  type: Twitter
  description: 'null'
- url: https://www.linkedin.com/company/toolhouseai
  name: Toolhouse LinkedIn
  type: LinkedIn
  description: 'null'
- url: https://discord.com/invite/xPvyBxhHtu
  name: Toolhouse Discord
  type: Discord
  description: 'null'
- url: https://youtube.com/@toolhouseai
  name: Toolhouse YouTube
  type: YouTube
  description: 'null'
- url: https://help.toolhouse.ai/
  name: Toolhouse Support
  type: Support
  description: 'null'
- url: https://toolhouse.betteruptime.com/
  name: Toolhouse Status
  type: StatusPage
  description: 'null'
- url: https://docs.toolhouse.ai/toolhouse/sitemap.md
  name: Toolhouse Documentation Sitemap
  type: Sitemap
  description: Markdown sitemap covering the entire docs.toolhouse.ai corpus including developer references, integrations,
    and capabilities pages.
- url: https://docs.toolhouse.ai/toolhouse/llms-full.txt
  name: Toolhouse LLMs.txt
  type: LlmsText
  description: Full Toolhouse documentation corpus packaged as a single LLM-friendly text file for AI agent ingestion.
- url: https://github.com/toolhouseai/toolhouse-sdk-python
  name: Toolhouse Python SDK
  type: SDKs
  description: Official Python SDK for integrating Toolhouse tools and agent invocations into Python applications (latest
    release v1.3.1).
- url: https://github.com/toolhouseai/toolhouse-sdk-typescript
  name: Toolhouse TypeScript SDK
  type: SDKs
  description: Official TypeScript/JavaScript SDK for Toolhouse integration.
- url: https://github.com/toolhouseai/client
  name: Toolhouse Client (JavaScript)
  type: SDKs
  description: Lightweight JavaScript library for Toolhouse AI agents / workers.
- url: https://github.com/toolhouseai/toolhouse-mcp
  name: Toolhouse MCP Server
  type: Integrations
  description: Connect Toolhouse tools to any MCP-compatible client or agent.
- url: https://github.com/toolhouseai/mcp-distributed
  name: MCP Distributed
  type: OpenSource
  description: Open source project for managing remote MCP servers at scale.
- url: https://github.com/toolhouseai/n8n-nodes-toolhouse
  name: Toolhouse n8n Integration
  type: Integrations
  description: n8n workflow automation nodes for Toolhouse worker invocation.
- url: https://github.com/toolhouseai/toolhouse-examples
  name: Toolhouse Examples
  type: SampleApplications
  description: Example implementations and demos for the Toolhouse platform (TypeScript).
- url: https://github.com/toolhouseai/fastlane-demo
  name: Fastlane Demo
  type: SampleApplication
  description: Demonstration project showing how Fastlane helps publishers monetize content with AI workers.
- url: https://github.com/toolhouseai/toolhouse-assessment
  name: Toolhouse Assessment
  type: Tools
  description: Web tool that helps companies assess their organizational readiness to deploy AI workers.
- url: https://github.com/toolhouseai/toolhouse-teleprompter
  name: Toolhouse Teleprompter
  type: SampleApplication
  description: Smart teleprompter application built on Toolhouse.
- url: https://github.com/toolhouseai/toolhouse-agenticlabs
  name: Toolhouse Agentic Labs
  type: SampleApplication
  description: Experimental agent / worker development lab from Toolhouse.
- url: https://docs.toolhouse.ai/toolhouse/developers/workers-api
  name: Toolhouse Workers API Documentation
  type: Documentation
  description: Developer documentation for the Toolhouse Workers API used to invoke deployed workers over HTTP with text and
    NDJSON streaming, stateful run IDs, and conversation continuation.
- url: https://docs.toolhouse.ai/toolhouse/developers/authentication
  name: Toolhouse Authentication Documentation
  type: Authentication
  description: Documentation covering API key creation, the HTTPBearer scheme, and the API key management surface inside the
    Toolhouse account settings.
- url: https://toolhouse.ai/blog/toolhouse-changelog-august-2025
  name: Toolhouse Changelog (August 2025)
  type: ChangeLog
  description: Most recent published changelog summarizing August 2025 platform updates including MCP Discovery, Creator Rewards,
    and Together AI model support.
- url: https://toolhouse.ai/blog/introducing-mcp-discovery-effortless-agent-superpowers-in-toolhouse
  name: Introducing MCP Discovery
  type: Article
  description: Announcement of MCP Discovery for effortless agent capability expansion through MCP servers.
- url: https://toolhouse.ai/blog/introducing-toolhouse-rag-free-yourself-from-rag-complexity
  name: Toolhouse RAG Announcement
  type: Article
  description: Announcement of Toolhouse RAG, the managed retrieval-augmented generation tool included with worker runs.
- url: https://toolhouse.ai/blog/connect-toolhouse-agents-to-zapier-in-minutes
  name: Toolhouse Zapier Integration Announcement
  type: Article
  description: August 2025 announcement of native Zapier integration for connecting Toolhouse workers to thousands of Zapier-enabled
    apps.
- url: https://toolhouse.ai/blog/now-running-on-together-ai-models
  name: Now Running On Together AI Models
  type: Article
  description: August 2025 announcement of Together AI model support across Toolhouse workers.
- url: plans/toolhouse-plans-pricing.yml
  name: Toolhouse Plans And Pricing
  type: Plans
  description: API Commons Plans 0.1 description of Toolhouse's published pricing tiers (Business, Business Max, Enterprise)
    reconciled against toolhouse.ai/pricing on 2026-05-22.
- url: rate-limits/toolhouse-rate-limits.yml
  name: Toolhouse Rate Limits
  type: RateLimits
  description: API Commons Rate Limits 0.1 description of Toolhouse's credit-based quotas, per-tier request defaults, schedule
    minimum interval, and attachment constraints.
- url: finops/toolhouse-finops.yml
  name: Toolhouse FinOps
  type: FinOps
  description: FOCUS-aligned FinOps framework definition mapping Toolhouse credit consumption and operational meters for cost
    allocation and unit economics.
- url: json-ld/toolhouse-context.jsonld
  name: Toolhouse JSON-LD Context
  type: JSONLD
  description: JSON-LD context defining Toolhouse domain vocabulary (AIWorker, Agent, WorkerRun, Tool, Schedule, ApiKey, McpServer)
    and linked-data mappings.
- url: json-schema/toolhouse-agent-schema.json
  name: Toolhouse Agent Schema
  type: JSONSchema
  description: JSON Schema describing the Toolhouse agent / worker entity model and related types.
- url: json-structure/toolhouse-agent-structure.json
  name: Toolhouse Agent Structure
  type: JSONStructure
  description: Structured documentation of the Toolhouse Agent / Worker object model.
- url: rules/toolhouse-rules.yml
  name: Toolhouse Spectral Rules
  type: SpectralRules
  description: Spectral ruleset enforcing Toolhouse API conventions and standards.
- url: vocabulary/toolhouse-vocabulary.yml
  name: Toolhouse Vocabulary
  type: Vocabulary
  description: Vocabulary and taxonomy for the Toolhouse AI worker / agent platform domain.
- name: Features
  type: Features
  data:
  - name: AI Worker Deployment from a Prompt (Agent Editor)
  - name: Agent Studio (v2)
  - name: Workers API (REST + Streaming + NDJSON)
  - name: Stateful Conversational Memory (X-Toolhouse-Run-ID)
  - name: Attachments via URL or Base64 (10 MB max)
  - name: MCP Discovery
  - name: MCP Server Integration
  - name: Toolhouse RAG
  - name: Skills and Knowledge (Templates, Guidelines, Reference)
  - name: Virtual Computer (Python Sandbox)
  - name: Browser Automation
  - name: Web and Social Media Search
  - name: Image Generation, Editing, and Vision
  - name: Document Parser
  - name: File Download
  - name: Memory
  - name: Scheduled Worker Runs (cron, 10-minute floor)
  - name: Worker Email Inbox
  - name: Worker Run Logs (incl. MCP Server Logs)
  - name: Worker Run Metrics And Volume Reporting
  - name: API Key Management
  - name: Organizations
  - name: Agent Subscriptions, Monetization, and Transfer
  - name: Streaming Responses (Text + NDJSON)
  - name: Free Unlimited Tokens On All Plans
- name: Use Cases
  type: UseCases
  data:
  - name: AI Worker Backend-as-a-Service
  - name: Deal Analysis and Market Research Automation
  - name: Document Processing (Policy Abstraction, Summarization, Query)
  - name: Knowledge Assistants For Policy and Business Data
  - name: Lead Generation and Outbound Automation (SMS, Cold Calling, Inbound)
  - name: Marketing and Content Creation (Listings, Copy, Image Generation)
  - name: Automated Worker Scheduling
  - name: Tool-Augmented LLM Pipelines
  - name: Agent Studio Chat
  - name: MCP Client Integration
  - name: Production Worker Deployment
- name: Integrations
  type: Integrations
  data:
  - name: OpenAI
  - name: Anthropic Claude
  - name: Together AI
  - name: Groq
  - name: NVIDIA (Llama 3.1 Nemotron 70B)
  - name: Tavily (Official Partner)
  - name: Vercel AI SDK
  - name: LlamaIndex
  - name: n8n (Toolhouse n8n Nodes)
  - name: Zapier
  - name: Supabase
  - name: Lovable
  - name: GitHub
  - name: GitLab
  - name: Bitbucket
  - name: Stripe
  - name: Square
  - name: QuickBooks
  - name: Salesforce
  - name: HubSpot
  - name: Microsoft Dynamics 365
  - name: Google Workspace (Sheets, Docs, Drive)
  - name: Microsoft Teams
  - name: Slack
  - name: Discord
  - name: Notion
  - name: MCP Clients (via MCP Server and MCP Discovery)
- name: MCP Server
  url: https://github.com/toolhouseai/toolhouse-mcp
  type: MCPServer
created: '2026-03-26'
modified: '2026-05-22'
position: Consumer
segments:
- AI Agents
- AI Workers
- Agent Infrastructure
description: 'Toolhouse is a Backend-as-a-Service platform for building, deploying, and managing "AI workers"

  (Toolhouse''s current product noun, used interchangeably with "AI agents"). The homepage frames the value

  proposition as "Turn your AI chats into affordable and reliable AI workers that do work while you''re

  busy," and the docs define a worker as "a system that carries out a task with three components: a trigger,

  a process that may include specialized skills, and tools or systems it can connect to." Builders describe

  a task in plain language in the Agent Editor and ship a worker to production; developers can do the same

  programmatically via SDKs. Every worker is automatically wired into Toolhouse''s pre-integrated

  capabilities, including web and social media search, scraping, Toolhouse RAG, persistent memory, the

  Virtual Computer code-execution sandbox, image generation/editing and vision, document parsing, file

  download, and MCP Discovery for connecting to thousands of MCP servers at runtime. The platform exposes

  a Workers API (agents.toolhouse.ai) for invoking deployed workers over REST with text and NDJSON

  streaming and stateful conversation continuity (X-Toolhouse-Run-ID), a Platform API for full worker and

  run lifecycle management, scheduling (10-minute minimum cadence), metrics, monetization, and integrations,

  plus no-code workflow integration through Zapier and Toolhouse n8n Nodes. Toolhouse runs on a broad model

  substrate including OpenAI, Anthropic, Together AI, Groq, and NVIDIA model offerings. The current public

  pricing surface is Business ($500/mo, 25,000 credits, 50 workers), Business Max ($1,200/mo, 80,000

  credits, 500 workers), and custom Enterprise — all plans include "free unlimited tokens" and a 14-day

  free trial. Toolhouse reports 7,000+ teams on the platform and 40+ integrations and is backed by

  NextGenerationEU funding.'
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com
specificationVersion: '0.19'