gotempsh/temps

Temps – Self‑hosted AI‑native PaaS

What it is – Temps is an open‑source, single‑binary platform you install on a server (or run in the cloud) that replaces a whole stack of SaaS tools. It lets you deploy apps from Git, run AI‑driven agents, monitor performance, capture errors, replay sessions, send transactional email, and more – all from one dashboard.

Key capabilities

• AI‑native operations – 440+ CLI commands are mirrored in the UI and exposed as skills that LLM agents (Claude, Codex, OpenCode, etc.) can invoke. Agents can deploy code, run migrations, inspect traces, or add domains without custom glue.
• AI chat – A read‑only chat interface that answers questions using your own telemetry (metrics, logs, traces, revenue). Write actions are opt‑in and require your confirmation.
• AI gateway – One OpenAI‑compatible endpoint that proxies requests to any provider you supply (OpenAI, Anthropic, Google Gemini, xAI). Keys stay encrypted on your server and usage (tokens, latency, cost) is logged.
• Web analytics & session replay – Funnel analysis, visitor tracking and rrweb‑based replay run entirely on‑premise.
• Uptime monitoring & alerts – Status timelines, health checks, and alerts for deploy failures, cert expiry, backups, etc.
• Sentry‑compatible error tracking – Drop‑in DSN replacement that groups errors, shows stack traces with source context and sends alerts.
• Request logs & proxy visibility – Full HTTP request logs, including identification of AI crawler traffic, served via Cloudflare’s Pingora engine with automatic TLS.
• Transactional email – Manage DKIM domains and send mail via the Temps SDK or plug in AWS SES, Scaleway, or any SMTP relay.
• OpenTelemetry hub – Collect distributed traces, metrics and structured logs in one place; no separate Grafana/Prometheus/Jaeger stack needed.
• AI sandboxes – Secure execution environments on Firecracker microVMs (or experimental libkrun “microsandbox”). Each sandbox gets its own kernel, preview URLs, password protection, and an SDK compatible with @vercel/sandbox.
• Git‑push‑to‑deploy & managed services – Push to a repo, Temps builds, deploys and provisions Postgres, Redis, MinIO S3, MongoDB, with zero‑downtime rollouts.
• Broad language support – Auto‑detects frameworks (Next.js, Vite, Go, Python, Rust, Java, .NET, NestJS, Dockerfile, etc.) or lets you bring your own Dockerfile.

How you get started

1. Run the one‑liner installer: curl -fsSL https://temps.sh/deploy.sh | bash. It installs Docker (if needed), spins up a PostgreSQL container, installs the Temps binary as a system service, and creates an admin account.
2. Open the console URL printed by the installer, sign in, and click Deploy sample app to verify Docker, image pulling and routing.
3. Use the UI or CLI (bunx @temps-sdk/cli …) to connect Git providers, add domains, enable databases, and invite teammates.

Why use it – According to the README, Temps can replace a dozen separate SaaS products (Vercel, PostHog, Sentry, Datadog, Pingdom, Resend, etc.) for essentially $0 if you self‑host, while also giving you a programmable AI layer that can act on your own observability data.

Links – Website, Docs, GitHub Discussions.