All the Elixir libraries I’ve been publishing recently and OpenPencil are part of the same vision. The idea is much bigger than just building another design harness.

IT'S HERE! This is Imp: declarative self-Improving language model programming in Elixir. It's a port of DSPy to the BEAM ecosystem. That means you get signatures, optimizers, agent loops, retrieval and more, all with the reliability and concurrency of the actor model and the elegance of Elixir. This started as a test of the models' ability to port things across languages, but it's become much more advanced, with Optimize-Anything support and full-fledged MCP and ACP adapters as well. Basically a whole kit of building blocks for anything you want to make. If you're interested in the DSPy paradigm, Elixir, text optimization, or weird code ideas in general, give it a look and tell me what you think! Imp is experimental, it's 0.5.0, it probably needs some expensive benchmarking to prove its optimizers really work. I'd love to get the kinks worked out before doing that haha. Imp is open source, MIT licensed, and hopefully easy to read. Given enough agents, all bugs are shallow, right? Now, without further ado: Imp! github.com/deepfates/imp

QuackDB 0.5.26 is out. Since the first package release, it evolved a lot. I've tested it under load and in many different consumers. @duckdb isn't just a database. It reads Parquet, CSV, JSON, Iceberg and Delta straight from S3, does full-text search, spatial and window functions, and hands back dataframes. QuackDB wraps it all in Elixir: DuckDB supervised in OTP, Ecto queries, DDL and DML as Elixir data. github.com/elixir-vibe/quack…

If coding agents will write most of our code, what happens with our communities and sense of ergonomics? How does it impact our compilers and tools?

Now it supports multiple backends, including Laya for local inference powered by Elixir Nx: github.com/dannote/jev_nx

Planned for the next release, along with support for more providers and better onboarding. Also, OpenPencil will never charge you per seat or force you to store your data in a specific cloud.

Jido Code Multiplayer (multi-actor)

I love the taste behind the choice of libraries Nuxt UI is built on. The cascade of CSS variables as design tokens, tailwind-variants for defining component styles, and Reka UI as a solid foundation make for a great, well-thought-out architecture for modern frontends. This approach is much more flexible than this demo can show. It’s a pleasure to watch the project evolve.

OpenPencil v0.15.0 is out ✏️ The biggest release yet: component variants and published libraries, a redesigned interface, guides and snapping, and an AI chat that finally works like a workspace. github.com/open-pencil/open-…

OpenPencil 0.15.1 just released. It’s good. As they say at a certain fruit company, it’s better than ever. I’m too tired to write a proper announcement post. I’ll do that tomorrow.

Redesigned the icon

Great to see Jev in ReqLLM. I've been playing with the model from the other end and ended up with a small GenServer-shaped client: github.com/dannote/jev Jev answers typed questions with probabilities, never text. On the BEAM that's just a message. So you reply to Jev from a callback, its answer arrives at 𝚑𝚊𝚗𝚍𝚕𝚎_𝚊𝚗𝚜𝚠𝚎𝚛/𝟹, and routing is pattern matching with confidence thresholds as guards. Crashed calls become error messages under a 𝚃𝚊𝚜𝚔.𝚂𝚞𝚙𝚎𝚛𝚟𝚒𝚜𝚘𝚛, recursion is a clause that replies again.

ReqLLM 1.24 is out Includes Jev support Includes LMStudio support Video, LiveBook and some helpers coming soon Jev is a whole new thing, so cool hex.pm/packages/req_llm

I’ve released a new version of OpenPencil, but immediately found a few regressions. A patch release will follow soon, and then I’ll make a proper announcement.

Impressive

WebMCP is about to land in OpenPencil. Just a couple more PRs left to merge before the new release. A lot of effort has gone into making both the UI and the internals more mature. I follow the competition closely, and I can see that as a product, OpenPencil is falling behind in terms of AI features. That is temporary. Unlike many competitors, who stopped at MVP-level Figma compatibility, I intended to build a strong foundation that would allow even heavy corporate workflows to migrate seamlessly to OpenPencil. The upcoming release is a big step in that direction. Internally, I’ve also managed to achieve close to 100% compatibility when reading and writing really complex files with extensive use of components and instances. This work still needs some polishing and refactoring, so I’ve delayed it until the next release. The self-hostable cloud is already in active development too. Expect more news soon.

Very interesting experiment: github.com/digitallyinduced/…

Pushed a few more fixes. After running it for a day, I can say that the abnormal usage drain has stopped.

We need that in Elixir too

By the way, I’m still maintaining my Pi setup: github.com/dannote/dot-pi

Try to install also this one, it should give your model more power for elixir. phxagents.dev/ I did also integration to Pi.