Ocsigen in 2026: what we're working on
What are we actually working on these days? Faster Wasm compilation, a new documentation toolchain, hierarchical module names, mobile through Capacitor, AI coding skills... Here is an honest look at the state of Ocsigen in 2026: what is in progress, what is on the roadmap and waiting for funding, and where you can lend a hand.
Ocsigen has always had a single, ambitious goal: to give the OCaml community a complete, top-quality suite of tools that makes Web and mobile programming extremely simple, while fitting naturally into the existing ecosystem of OCaml libraries. From the compiler to the browser, from the server to the database, everything should work together, be type-safe end to end, and stay at the best level of what the state of the art has to offer.
Here is a snapshot of what we are currently working on, and where we would love some help.
Work in progress
A lot is happening around compilation to the Web. We are pushing wasm_of_ocaml further: performance improvements, a new syntax for hand-written Wasm, and a systematic comparison between the js_of_ocaml and wasm_of_ocaml runtimes to detect and fix bugs.
On the documentation side, we are building wodoc, an odoc driver that can produce a complete website: manual, API reference, blog, and a Markdown version meant for AI agents. This goes hand in hand with a modernisation of our documentation infrastructure (moving ocsigen.org over to wodoc) and a broad content refresh across all projects.
The core framework is getting a more polished surface. Across the server, Eliom, the Toolkit and Ocsigen Start, we are moving to hierarchical module names (for instance Eliom.Service instead of Eliom_service), and factorising the build system into the shared ocsigen-dune-rules. On mobile, Ocsigen Start is switching from Cordova to Capacitor. Ocsipersist is gaining type-safe persistent references built on Deriving. The Toolkit is getting new reactive widgets, contributed by Be Sport. And we are writing AI skills for agent-assisted coding with Eliom and Start.
Backlog — looking for funding
We have a lot of ideas for the next steps, and several of them are ready to start as soon as funding allows:
- Lwt: reaching the performance of effect-based libraries, and multi-domain compatibility.
- Ocsipersist: adding a fast, 100% OCaml backend.
- Eliom: off-line apps out of the box, client-side-only use, and distributed apps backed by an Irmin-like database.
- Toolkit: documenting how to use it for client-only
js_of_ocaml/wasm_of_ocamlapps. - Server: running as a plain executable, with no configuration file.
- Apps: a collaborative editor, and a distributed content management system, ...
Features we need in neighbouring projects
Some of our goals depend on improvements elsewhere in the ecosystem (all currently unfunded): the missing odoc features that would let us drop wodoc's hacks, complete multi-tier programming support in Merlin/LSP, and the same in Dune so we can avoid dirty build-system workarounds.
How you can help
- Contribute.
- Fund the projects you care about.
- Tell us about your needs and feature requests.
- Help us spread the word.
Thanks
A huge thank you to all our funders, and especially to Jane Street, Tarides, OCaml Software Foundation, NLnet and Ahrefs.