Skip to main content
Squasher connects software reliability and AI product quality in one evidence trail. Capture errors, logs, traces, uptime, and replay alongside AI inputs, outputs, tool calls, token usage, cost, and user feedback—then investigate, fix, and verify what users experienced.

​
What Squasher gives you

  • AI triage on every new error group so you get a summary, severity, and fix direction faster.
  • SDK-based error capture for application-level context.
  • Log drains for infrastructure and platform errors when you do not want to install an SDK.
  • OpenTelemetry-native AI generation traces with prompt, response, tool, token, cost, and feedback lineage.
  • Exportable scored evidence for evaluation and training datasets.
  • Alerting, status monitoring, and hosted fix workflows in the same product.

​
How dashboard work fits together

Each dashboard page shows its organization and project scope. Keep that scope fixed while you move through one response workflow:
  1. Setup creates an ingest credential, connects one runtime or telemetry path, and verifies a real signal.
  2. Explore moves between Logs, Traces, and Sessions while it keeps the compatible time, environment, service, and search scope in the URL. A selected event, trace, span, or session opens beside the result list and remains available through reload, shared links, and browser history.
  3. Issues groups recurring runtime failures into a triage queue with an owner, priority, automation state, and next action.
  4. Incidents coordinates active response through separate Response, Evidence, Resolution, Communications, and Follow-ups views.
  5. Project settings controls the response policy, notifications, incident policy, data, and credentials for that project.
This flow keeps observed evidence separate from internal response and customer publication. An internal issue or incident does not become a public status update unless a publication rule or a responder explicitly publishes it.

​
Start with the right path

Quickstart

Set up Squasher in a Next.js app, Node.js service, or provider log drain.

Agent DX

Give coding agents the docs, API, CLI, and MCP entry points they should use.

Next.js SDK

Capture route, server, and client errors with framework-aware helpers.

Node.js SDK

Instrument Express, Fastify, Hono, Bun, or plain Node.js services.

Vercel log drain

Start with zero-code capture for deployment and runtime logs.

​
Supported paths

​
For agents

Start with Agent DX or https://docs.squasher.ai/llms.txt. Use this routing:
  • Need to change app code: pick the SDK guide for the runtime, initialize early, and send one verification event.
  • Need live Squasher data: connect the MCP server, call search, then call execute for the exact operation.
  • Need repeatable terminal commands: use the CLI setup guide and include --project <project_id>.
  • Need custom integration code: use the API reference and OpenAPI contract.

​
Next steps

  1. Follow the quickstart.
  2. Configure the SDK or integration that matches your runtime.
  3. Review troubleshooting if your first events do not appear.