Can I use Bicho without an LLM or AI agent? #72
|
Can a script or a person investigate Bicho errors without connecting an LLM? Which interface should it use? |
Replies: 1 comment
|
Yes. The authenticated HTTP API supports investigation without an LLM or MCP client. Use ordinary HTTP requests from a trusted script or server to list projects, search issues and events, fetch captured JSON, read status history, obtain statistics, and inspect processing health. Bicho has no built-in dashboard. You can use For a basic read-only workflow, request HTTP also exposes some inputs that the MCP tools omit, including event/statistics status filters and history time filters. Collection through an SDK or compatible ingestion transport is separate from how you investigate the stored evidence. Neither capture nor API access requires model inference. See the HTTP API reference and MCP comparison. |
Yes. The authenticated HTTP API supports investigation without an LLM or MCP client. Use ordinary HTTP requests from a trusted script or server to list projects, search issues and events, fetch captured JSON, read status history, obtain statistics, and inspect processing health.
Bicho has no built-in dashboard. You can use
curl, write a small reporting script, or build a separate interface around the API. Keep the admin token on a trusted server or in secure client storage. Publishing it in browser code would expose administrative access; CORS does not make an exposed token safe.For a basic read-only workflow, request
GET /api/projects, choose a returned project ID, then requestGET /api…