
LangChain
Agent engineering platform for building LLM applications from composable parts
128+ repos to power your next build

Agent engineering platform for building LLM applications from composable parts

Data-backed skills and practices for solo founders—ground your AI workflows in how one-person companies actually win.

One Person Company agent team patterns—delegate research, building, and ops to coordinated AI helpers.

Multi-agent operating system concepts for one-person companies—coordinate AI roles around real founder workflows.

Modern Python API framework—the default backend for AI services that need speed, typing, and OpenAPI docs.

Data and AI app UI in pure Python—launch internal tools and MVPs without splitting into a separate frontend repo.

Ship ML demos in Python minutes—wrap models in shareable UIs for users, clients, and investors without frontend drama.

Structured LLM outputs into Pydantic models—stop parsing Markdown guesswork and ship reliable JSON features.

Eval and red-team your prompts—catch regressions and unsafe outputs before customers do.

Programmatic prompting framework—optimize LM pipelines with code and metrics instead of endless hand-tuned strings.

Conversational data analysis for non-experts—ask questions about datasets without living in notebooks all day.

Text-to-SQL assistant framework—let users query databases in plain language for internal tools and demos.

Lifelong-learning agent famous in Minecraft—study open-ended skill acquisition ideas for ambitious agent products.

Orchestrate many expert models via an LLM planner—route tasks to the right model instead of one generalist call.

Desktop computer-use agent framework—experiment with AI that clicks GUIs when APIs are missing.

Minimal task-loop agent that taught autonomy basics—learn memory and prioritization in a tiny codebase.

Open Notion-style workspace—notes, docs, and boards you can self-host when proprietary wiki lock-in worries you.

Publishing platform for creator businesses—memberships, newsletters, and blogs in one indie-friendly stack.

Classic open headless CMS—give non-technical editors an admin while your indie app consumes clean APIs.

TypeScript headless CMS loved by Next.js solos—manage content with code-first schemas and a polished admin.

Open feature flags and experimentation—ship AI features gradually and learn with A/B tests you control.

Simple self-hosted analytics—own your metrics with a clean UI and without renting another surveillance suite.

Privacy-friendly web analytics that stays simple—see traffic without creepy tracking theater or heavy scripts.

Open product analytics with session tools—understand user behavior without sending every event to a closed black box.