asteroid
Asteroid lets you build, run, and manage AI browser agents that automate real web workflows — filling forms, navigating portals, extracting data, and…
What it can do
- GetContext: Get your current user context including organization IDs and names. Call this first to look up an organization's UUID by its name before using other tools. The Asteroid platform is at http
- AgentProfilesList: List agent profiles for an organization. Profiles are reusable browser environments — NOT agents. Each profile can configure: proxy routing (country-specific), automatic captcha sol
- AgentProfilesCreate: Create an agent profile — a reusable browser environment config (not an agent). Profiles can include proxy settings, captcha solving, cookie/cache persistence, and vault credentia
What data it sees
Do you need an account
No: the server works without sign-in
Asteroid lets you build, run, and manage AI browser agents that automate real web workflows — filling forms, navigating portals, extracting data, and completing multi-step tasks on any website.
With the Asteroid connector, Claude becomes your agent operations console:
🛠️ Build agents
Create agents and define workflows in plain conversation. Claude assembles the workflow graph, validates output schemas, picks the right node tools, and publishes versions when you're ready.
▶️ Run & steer executions
Start executions, watch their activity step by step, pause/resume/cancel, and send mid-run messages to steer an agent while it works.
🌐 Manage browser environments
Configure reusable agent profiles with country-specific proxies, automatic captcha solving, persistent cookies/cache, and securely vaulted credentials that agents reference without ever seeing the secrets.
📄 Inspect results
Pull execution results, produced files, and persistent shared files straight into your Claude conversation for analysis.
How to connect
How to connect asteroid to Claude
The server address to paste into Claude settings:
https://asteroid--asteroid.run.tools
- Open Claude (claude.ai or the desktop app).
- Go to Settings → Connectors.
- Click Add custom connector.
- Paste the server address copied below into Remote MCP server URL and click Add.
- In a chat, click + → Connectors and switch the new connector on.
Custom connectors are available on Free, Pro, Max, Team and Enterprise plans (Free is limited to one). On Team and Enterprise an organization Owner adds the connector first under Organization settings → Connectors.
Authorization
After you click Add, a sign-in window for the service opens. Sign in with your own account and approve access. Claude never sees your password.
How to connect asteroid to ChatGPT
The server address to paste into ChatGPT settings:
https://asteroid--asteroid.run.tools
- Open ChatGPT in a browser (chatgpt.com). A Plus, Pro, Business, Enterprise or Edu plan is required.
- Turn on developer mode once: Settings → Apps → Advanced settings → Developer mode.
- Open Settings → Connectors and click Create.
- Fill in the form: Name (anything), Description (one line about what the service does), MCP server URL (copy it below).
- Under Authentication choose OAuth if the service requires sign-in, otherwise None. Click Create.
- In a new chat open + → Apps/Connectors and enable the connector.
OpenAI has renamed this section before (Connectors → Apps/Plugins). If the label differs, search settings for "developer mode". On Business/Enterprise workspaces an admin must allow custom connectors first.
Authorization
On first use ChatGPT opens the service's sign-in window. Sign in and approve access.
How to connect asteroid to Cursor
The server address to paste into Cursor settings:
https://asteroid--asteroid.run.tools
Fastest: click Open in Cursor below and confirm the prompt.
Manually:
- In Cursor open Settings → Cursor Settings → MCP and click Add new global MCP server.
- Paste the JSON copied below into
~/.cursor/mcp.json(per project:.cursor/mcp.jsonin the repo root). - Save the file. The server appears in the MCP list; authorize it there if asked.
Authorization
If the service needs sign-in, an authorize button appears next to the server in the MCP list.
Server tool list (31)
Raw names from tools/list. Only developers need these.
| getContext | Get your current user context including organization IDs and names. Call this first to look up an organization's UUID by its name before using other tools. The Asteroid platform is at https://platform.asteroid.ai — use this as the base for any dashboard links. |
| agentProfilesList | List agent profiles for an organization. Profiles are reusable browser environments — NOT agents. Each profile can configure: proxy routing (country-specific), automatic captcha solving, persistent cookies/cache across executions, and securely stored credentials (passwords, TOTP/2FA secrets) referenced in agent instructions via ##VARIABLE_NAME## syntax. |
| agentProfilesCreate | Create an agent profile — a reusable browser environment config (not an agent). Profiles can include proxy settings, captcha solving, cookie/cache persistence, and vault credentials for authentication. |
| agentProfileGet | Get an agent profile by ID, including its proxy, captcha, persistence, and credential settings. |
| agentProfileUpdate | Update an agent profile's settings (proxy, captcha solving, cookie/cache persistence, credentials). |
| agentProfileDelete | Delete an agent profile. |
| agentProfileClearBrowserCache | Clear the persisted browser cache (cookies, storage) for an agent profile, resetting its browser state. |
| agentList | List all agents for an organization. Pass the `organizationId` from `getContext` (use `getContext` first if you only know the org name). |
| agentCreate | Create a NEW agent (not a new version of an existing one). Requires `organizationId` (from `getContext`), a human-friendly `name`, and an initial `workflow` graph — the backend always seeds an agent with a first workflow. For a blank-but-runnable agent, pass a minimal starter: one start `iris` node plus `success` and `failure` `output` nodes, with `settings.new_agent_loop: true`. For a pre-built agent, pass a fuller graph (same JSON shape as `workflowCreate`). Use `listAvailableTools` to pick node tools and `schemaValidate` for any output schemas. The created agent's first workflow is unpublished; publish it with `workflowPublish` before `agentExecutePost`. |
| agentExecutePost | Start an execution for the given agent |
| agentDelete | Permanently delete an agent and all of its workflows by `agentId`. This is irreversible. Only allowed when all of the agent's executions are in a terminal status (completed, cancelled, or failed); otherwise the request is rejected. Use `agentList` to find the `agentId` first. |
| executionsList | List executions with filtering and pagination |
| executionGet | Get a single execution by ID with all details |
| executionActivitiesGet | Get activities for an execution |
| executionAgentFilesGet | Get all files for an execution, grouped by directory (uploads, downloads, workspace, shared, debug). Files are tracked by the file syncer daemon during execution. Debug files are only visible to admins. |
| executionContextFilesGet | Get all context files attached to an execution |
| executionStatusUpdate | Update the status of an execution (pause, resume, or cancel) |
| executionUserMessagesAdd | Add a user message to an execution |
| listAvailableTools | List all available tools/capabilities that can be assigned to AI Task nodes in workflows |
| workflowsList | List all versions of an agent (lightweight refs with IDs, parents, and version numbers). |
| workflowCreate | Create a new agent version with a full graph (nodes, transitions), rules, and settings. This is how you update an agent's behavior. Set `parentId` to the current version's ID to maintain lineage. The new version is unpublished until you call `workflowPublish`. |
| workflowGet | Get a full agent version by ID, including the complete node graph, rules, and settings. |
| workflowPublish | Publish an agent version, assigning it a permanent version number and making it the default for `agentExecutePost`. |
| workflowExecute | Execute a specific agent version (published or unpublished) directly, useful for testing before publishing. |
| workflowSyncExecution | Sync a running execution to use a specific agent version. |
| schemaValidate | Validate a JSON schema against OpenAI structured output requirements |
| agentSharedFilesGet | List all persistent shared files for an agent. These files persist across executions and are automatically restored when a new execution starts. |
| agentSharedFilesDelete | Delete a file from the agent's persistent shared storage. |
| agentSharedFilesDeleteAll | Delete all shared files for an agent. |
| agentSharedFilesFreeze | Set the frozen status of a shared file. Frozen files are preserved across executions and cannot be overwritten by the agent. |
| searchDocs | Search Asteroid documentation for guides on building agents, node types, and best practices |