HereGo.
Documentation

HereGo docs.

Connect AI agents over Model Context Protocol, or integrate directly with the REST API. MCP is the fastest path — it is a stateless adapter over the same REST endpoints, so everything you can do in REST is one tool call away.

MCP-FIRST · RECOMMENDED

MCP Server

https://herego.video/mcp

The preferred way for AI agents to request visual evidence. Three tools over Streamable HTTP — no auth, no account, no API key.

  • request_video_evidence
  • get_video_evidence
  • delete_video_request
Explore the MCP Server→
REST

REST API

https://herego.video/v1

Direct HTTP integration for applications: create requests, poll status, and stream submitted videos with request-scoped tokens.

  • POST /v1/requests
  • GET /v1/requests/{id}
  • GET /v1/videos/{id}
Read the REST API Docs→
Quickstart

Get evidence in minutes.

Follow the agent path over MCP, or integrate directly over REST. Both start the same way: create a request, share the link, retrieve the video.

1

Point your MCP client at HereGo

No account, no API key, no MCP-level authentication. Add the endpoint to your MCP client configuration:

https://herego.video/mcp
Client configuration
{
  "mcpServers": {
    "herego": {
      "url": "https://herego.video/mcp"
    }
  }
}
2

Generate a video submission request

Call request_video_evidence with an instruction for the human. This creates the request — it does not record anything yet.

Example tool call
{
  "prompt": "Please show me what happens when you try to check out.",
  "expires_in_hours": 24
}
Example result
{
  "id": "req_abc123",
  "status": "awaiting_submission",
  "submission_url": "https://herego.video/r/abc123",
  "access_token": "rat_9f8b7c6d5e4f3a2b1c0d9e8f7a6b5c4d",
  "expires_at": "2026-09-08T14:30:00Z"
}
3

Share the link, then poll for the video

Send the submission_url to the human and keep the access_token to yourself. Then call get_video_evidence until the status flips to submitted — the actual video arrives as MCP media content.

Example tool call
{
  "request_id": "req_abc123",
  "access_token": "rat_9f8b7c6d5e4f3a2b1c0d9e8f7a6b5c4d"
}
Example result
{
  "request_id": "req_abc123",
  "status": "submitted",
  "expires_at": "2026-09-08T14:30:00Z",
  "video": {
    "mime_type": "video/webm",
    "size": 482113
  }
}
// + the actual video attached as MCP media content
awaiting_submissionWaiting on the human — poll again later
submittedVideo attached as MCP media content
expiredGone — create a new request instead
4

Delete the request when you're done

Call delete_video_request with the same request_id and access_token to remove the entire request and its video.

Example tool call
{
  "request_id": "req_abc123",
  "access_token": "rat_9f8b7c6d5e4f3a2b1c0d9e8f7a6b5c4d"
}
Example result
{
  "request_id": "req_abc123",
  "deleted": true
}

GET /api

Machine-readable discovery →

openapi.json

OpenAPI specification →

llms.txt

Agent quickstart →

EARLY ACCESS

Join the waitlist to access the Agent Management Dashboard.

Or use the MCP or APIs for free — no sign-up or API key required.