Superlog

Superlog is an open-core observability workspace built on OpenTelemetry, developed by Pulsent Labs Inc. and backed by Y Combinator (P26 batch). It ingests traces, logs, and metrics from your services, groups noisy signals into clear incidents, and uses AI agents to prepare resolution pull requests — all triggered by a single coding-agent prompt. The core community edition is licensed under Apache 2.0 and self-hostable.

What It Is

Superlog sits in the observability category but extends beyond passive monitoring. Rather than surfacing raw logs and leaving engineers to triage, it runs an agentic pipeline: an open-source "wizard" explores your codebase, instruments it with OpenTelemetry spans, metrics, and structured logs via a generated PR, then continuously scans for new failure modes. When errors occur, a fingerprinting engine groups similar signals into a single incident with a severity score (SEV1–3) and an impact assessment, and an agent prepares a fix PR. The repository contains the web app, API, OTLP ingest proxy, worker processes, Postgres schema, ClickHouse-backed telemetry queries, and pluggable agent runner interfaces.

How the Agentic Workflow Operates

The install path is intentionally prompt-driven: running npx skills add superloglabs/skills --all in a coding agent (such as Cursor or Claude) triggers the wizard to explore the codebase and open an superlog/install-otel PR. After that, Superlog operates continuously:

• Continuous scan — detects new alert conditions, metrics, and dashboards as the codebase evolves, preventing "observability decay."
• Incident grouping — fingerprinting merges repeated error logs into one incident with a summary, severity, and impact statement.
• Resolution PRs — for each incident, an agent prepares a fix PR; if a "Confidence Gate" check fails, it posts findings to Slack and tags relevant engineers.
• MCP surface — logs, traces, metrics, alerts, and dashboards are all accessible through MCP, so agents can query and create dashboards without a separate UI workflow.

Architecture and Open-Source Model

The GitHub repository (superloglabs/superlog) is structured as a monorepo with TypeScript throughout:

• apps/web — Vite/React frontend
• apps/api — HTTP API
• apps/proxy — OTLP intake proxy
• apps/worker — background workers and agent orchestration
• packages/db — Drizzle schema and migrations
• packages/fingerprint — telemetry fingerprinting helpers

Prerequisites for self-hosting are Node.js 20+, pnpm 9+, and Docker. The community edition ships a default agent runner that records a local incident summary. Superlog Cloud is the hosted edition with a free tier and metered plans.

Current Status

The repository was created in June 2026 and shows active commit activity. As of the latest data, the project has 640 stars and 35 forks on GitHub, with 8 open issues. The project is tagged as a Y Combinator P26 company. The hosted Superlog Cloud product is live alongside the self-hosted community edition.