This project is scheduled for launch
Launch date: Wednesday, September 30, 2026 at 08:00 AM UTC

Spanlens is an open-source (MIT) LLM observability platform that provides comprehensive monitoring and optimization for your large language model applications. It logs every OpenAI, Anthropic, and Gemini request with a single line of code, helping developers and teams track cost, latency, tokens, and overall performance.
Targeted at engineering teams and data scientists, Spanlens helps optimize LLM workflows, reduce operational costs, and ensure the reliability and security of AI-powered features.
Spanlens is essential for monitoring LLM API calls, identifying cost spikes, performance bottlenecks, and errors. It optimizes complex multi-step agent workflows by pinpointing slow or expensive steps. Product teams leverage Evals and Experiments to objectively compare prompt versions and models, ensuring quality improvements and cost efficiency before shipping to production. It also provides user analytics for per-user cost and volume tracking.
Spanlens offers a freemium model. A free tier includes 50,000 requests/month, 14-day retention, and one seat. Paid plans (Pro, Team, Enterprise) offer increased volumes, longer retention, more seats, and advanced support. Self-hosting is free forever, providing unlimited retention and full data control.
With a "one line of code" setup, Spanlens integrates seamlessly with minimal latency overhead (p99 under 3ms) due to asynchronous ingestion. It provides comprehensive documentation and community support, with higher tiers offering email, Slack, webhooks, and dedicated support. Data remains secure with PII detection, auto-masking, and self-hosting options for complete privacy.
Built with TypeScript and Python, Spanlens supports Next.js, Node, and Edge environments. It integrates with major LLM providers like OpenAI, Anthropic, Google, and Mistral, as well as popular SDKs such as LangChain, LlamaIndex, and Vercel AI SDK. It's self-hostable via Docker Compose or a single binary and supports OpenTelemetry for existing tracing infrastructures.
Spanlens delivers a powerful, open-source solution for LLM observability, enabling teams to monitor, optimize, and secure their AI applications efficiently. Its rich feature set, from real-time logging and cost tracking to advanced agent tracing and prompt experimentation, makes it an indispensable tool for any organization leveraging large language models. Start with Spanlens today to gain complete visibility and control over your LLM operations.
Comments will be available once the project is launched.
HAESEONG JEON
Scraper Email for Product Hunt Daily leaderboard and peerlist, tinylaunch, uneed.
Get your brand featured here