jerrycan vs FastAPI
Hand-written Python vs AI-built Rust
FastAPI and jerrycan are both open-source frameworks you fully own — the honest difference is who does the building and what comes in the box. FastAPI optimizes for a human writing Python; jerrycan optimizes for an AI agent assembling a complete product. Here's the full read.
The short answer
Choose FastAPI if
- Your team writes Python daily and wants full manual control.
- You depend on the Python ecosystem: ML libraries, data tooling, existing code.
- You prefer picking your own auth, ORM and deployment stack piece by piece.
Choose jerrycan if
- An AI agent does your building and you want batteries it cannot wire wrong.
- You want compiled-Rust performance without writing Rust yourself.
- You want auth, data, files and payments in one coherent codebase from day one.
What is FastAPI?
FastAPI is a modern Python web framework for building APIs, built on standard Python type hints. You define your data shapes with Pydantic models, write endpoint functions, and FastAPI gives you validation, serialization and interactive API docs automatically. It's one of the most loved frameworks in the Python ecosystem, and deservedly so.
Where FastAPI shines
As a hand-writing experience, FastAPI is close to the state of the art: the type-hint-driven design catches errors early, the auto-generated docs are genuinely useful, performance is excellent for Python, and the surrounding ecosystem — every ML library, every data tool, decades of packages — is the largest in backend development. The documentation is a masterclass. If a human is typing the code, FastAPI is the default to beat.
Where FastAPI falls short
FastAPI's honest limitation is that it's a framework, not a product kit — and it assumes a skilled human doing sustained work. Auth, database migrations, file storage, payments and deployment are all yours to choose, wire and maintain from a sprawling menu of third-party options; the 'weekend of glue' is real, and every project re-litigates the same stack decisions. Python's interpreter and GC put a ceiling on performance and a floor on memory that compiled languages don't have. And for AI agents, all that flexibility is a liability: with twelve valid ways to wire anything, generated codebases drift into inconsistent choices — there's no single structure an agent can't get wrong.
What is jerrycan?
jerrycan is a self-hostable Rust backend framework for REST APIs, designed for a different builder: an AI agent. You describe the product; your agent assembles auth, data, files and payments from jerrycan’s built-ins as real code in your repository, then deploys it to a server you choose. The output is a compiled binary you own outright — no platform between you and your users, no meter on your growth.
Where jerrycan beats FastAPI
- Batteries: auth, data, files and payments are framework built-ins — no third-party stack assembly, no glue weekend.
- AI-buildability: one coherent structure designed so an agent composes correctly instead of choosing among twelve options.
- Performance envelope: compiled Rust — no interpreter at runtime, small binaries, fast cold starts, lower memory.
- Compile-time safety: the type system rejects a whole class of mistakes (human or AI) before anything ships.
- Idea to live product: one conversation including deployment, versus days-to-weeks of hand-wiring.
| FastAPI | jerrycan | |
|---|---|---|
| What it is | A Python web framework — you design and wire everything around it. | A Rust backend framework with batteries in — auth, data, payments, security. |
| Who writes the code | You do, endpoint by endpoint. | Your AI does — you describe the product and review the result. |
| Performance profile | Fast for Python, with an interpreter and GC in the loop. | Compiled Rust — small footprint, no interpreter at runtime. |
| Ownership | Fully open source — yours. | Fully open source — yours. On this one we agree. |
| Idea → live app | Days to weeks of hand-wiring auth, storage, deployment. | One conversation — built and deployed from your agent chat. |
Own the backend behind your SaaS
Point your AI at jerrycan — it builds a real Rust backend for your REST APIs that's yours to keep.