ErrorDX Documentation

ErrorDX is a hosted error-monitoring service. It ingests events from the Sentry SDKs you already use, groups them into issues, and gives you a dashboard for triage, alerts and team workflows — with nothing to install or operate beyond your app's SDK.

What you getCopy Anchor Link

• Sentry-compatible ingestion via /api/:projectId/store/ and /api/:projectId/envelope/.
• Project-based DSNs generated from your public keys.
• Issue grouping with event history and a searchable activity feed.
• Traces: a trace list with a span waterfall you can open span by span (attributes, status, self time), and p50/p95 latency per transaction.
• Structured logs on the same timeline as errors, with typed attributes.
• AI usage rolled up from gen_ai.* span attributes: calls and tokens per model, plus the tools and agents your agents reached for.
• A project overview per project — volume by severity, crash-free sessions, recent issues and the slowest transactions — and per-project storage accounting, so retention is a decision you can make with numbers.
• Alerts over email (SMTP/Plunk), Slack, Discord and webhooks, with retries and cooldowns.
• Source-map symbolication and release/artifact-bundle uploads via sentry-cli.
• Agent-first surfaces: an MCP server, a JSON CLI, scoped agent keys and an OpenAPI document.
• Per-organization plans with quota reporting in the dashboard.

Quick pathsCopy Anchor Link

How it is put togetherCopy Anchor Link

• Server (errordxd): Rust (Axum) with SQLite storage — REST API, Sentry-compatible ingestion, MCP, and in-process workers for alerts, symbolication and retention.
• Auth sidecar (auth): Bun process serving Better Auth (/api/auth/*) and Stripe billing endpoints against the same database.
• Storage: S3-compatible object storage for source maps and debug files.
• Web: the dashboard at app.errordx.com, a TanStack Start app over the REST API.

Why teams use itCopy Anchor Link

• Sentry-compatible: keep your existing SDKs, DSN format and sentry-cli workflows.
• Nothing to run: no servers, no database, no retention or upgrade chores.
• Agent-first: an MCP server, a JSON CLI and agent keys so your own agents can query issues and traces directly, backed by an OpenAPI document.