# OpenHelm

> OpenHelm turns high-level goals into scheduled, self-correcting AI agent jobs. Developers and AI agents drive it over a REST API (/v1) or via remote MCP servers (ChatGPT, Claude, Cursor). It runs work that takes minutes, hours, or recurs on a schedule, in isolated cloud sandboxes, and returns structured, evidence-backed JSON.

## Start here
- [MCP & API hub](https://openhelm.ai/mcp): MCP servers + REST API overview
- [API quickstart](https://openhelm.ai/docs/api-quickstart): run a task in one call
- [Live OpenAPI spec](https://api.openhelm.ai/v1/openapi.json): machine-readable API reference
- [llms-full.txt](https://openhelm.ai/llms-full.txt): the full developer docs as clean markdown

## Documentation
### Getting Started
- [Installation](https://openhelm.ai/docs/installation)
- [Quick Start](https://openhelm.ai/docs/quick-start)
- [Requirements](https://openhelm.ai/docs/requirements)

### For Developers
- [API Quickstart](https://openhelm.ai/docs/api-quickstart)
- [Authentication & API keys](https://openhelm.ai/docs/authentication)
- [Runs & the result envelope](https://openhelm.ai/docs/runs-and-results)
- [Long-running & recurring jobs](https://openhelm.ai/docs/long-running-jobs)
- [Connections API](https://openhelm.ai/docs/connections-api)
- [MCP servers](https://openhelm.ai/docs/mcp-servers)
- [Pricing & credits](https://openhelm.ai/docs/pricing-credits)
- [Deep research guide](https://openhelm.ai/docs/guides/deep-research)
- [SEO & GEO guide](https://openhelm.ai/docs/guides/seo-geo)
- [Email outreach guide](https://openhelm.ai/docs/guides/email-outreach)
- [Dev changes guide](https://openhelm.ai/docs/guides/dev-changes)

### Core Concepts
- [Projects & Goals](https://openhelm.ai/docs/concepts/projects-and-goals)
- [Jobs](https://openhelm.ai/docs/concepts/jobs)
- [Runs](https://openhelm.ai/docs/concepts/runs)
- [Fair Source](https://openhelm.ai/docs/concepts/fair-source)

### Scheduling
- [Schedule Types](https://openhelm.ai/docs/scheduling/schedule-types)
- [Manual Triggers](https://openhelm.ai/docs/scheduling/manual-triggers)

### Features
- [AI Planning](https://openhelm.ai/docs/features/ai-planning)
- [Self-Correction](https://openhelm.ai/docs/features/self-correction)
- [Memory System](https://openhelm.ai/docs/features/memory-system)
- [Data Export](https://openhelm.ai/docs/features/data-export)

### Reference
- [App Settings](https://openhelm.ai/docs/reference/settings)
- [Troubleshooting](https://openhelm.ai/docs/reference/troubleshooting)

## MCP servers (remote, Streamable HTTP)
- OpenHelm MCP (General-Purpose) — https://mcp.openhelm.ai/main/mcp: Spin up a virtual computer for any custom task — browse, run code, read data, and return structured results. → https://openhelm.ai/mcp/general-purpose
- OpenHelm Deep Research MCP — https://mcp.openhelm.ai/research/mcp: A virtual research environment that reads SEC filings, earnings prints, and live market data — grounded in official sources. → https://openhelm.ai/mcp/deep-research
- OpenHelm SEO & GEO Organic Growth MCP — https://mcp.openhelm.ai/seo/mcp: A virtual SEO environment — audit rankings, draft content, and measure AI engine citations of your brand. → https://openhelm.ai/mcp/seo-geo
- OpenHelm Email Outreach MCP — https://mcp.openhelm.ai/email/mcp: A virtual inbox your agent controls — send, monitor replies, and run the whole conversation to completion. → https://openhelm.ai/mcp/email-outreach
- OpenHelm Dev Changes MCP — https://mcp.openhelm.ai/dev/mcp: A virtual coding environment — clone the repo, make the change, run tests, and open a PR. → https://openhelm.ai/mcp/dev-changes
- OpenHelm Phone Reach-Out MCP — coming soon: A virtual phone your agent uses to call, converse, and act on the outcome — coming soon. → https://openhelm.ai/mcp/phone-reachout
- OpenHelm WhatsApp Reach-Out MCP — coming soon: A virtual WhatsApp inbox your agent uses to message, handle replies, and achieve the goal — coming soon. → https://openhelm.ai/mcp/whatsapp-reachout
- OpenHelm Online Buying MCP — coming soon: A virtual shopper that sources products, compares options, and checks out within your guardrails — coming soon. → https://openhelm.ai/mcp/buying-online

## Key facts for agents
- Auth: `Authorization: Bearer oh_live_...` (mint under Settings → API & Webhooks).
- Run a task: `POST /v1/runs` with `prompt` (+ optional `output_schema`, `model`). Returns 202 + `task_id`.
- Get the result: poll `GET /v1/runs/{id}` or supply a `callback_url` webhook (HMAC-signed).
- Recurring jobs fire the webhook on every run — register once, receive forever.
- Credentials: `POST /v1/connections` (token / mcp) or `POST /v1/connections/links` (OAuth providers).
- A run needing a missing connection fails with `failure_reason: "auth_required"` — never fabricated.
- Pricing: usage draws from plan credits (1 credit = $0.05); each run returns a cost estimate.