May 6, 2026
C you in the comments
Show HN: Adam – An embeddable cross-platform AI agent library
This tiny AI toolkit wowed coders — then sparked a ‘why on earth is this in C?’ pile-on
TLDR: Adam is a new all-in-one AI library built in C that aims to work almost everywhere, from desktops to phones to the web. The big community reaction wasn’t just “cool tool” — it was a loud, funny argument over why anyone would build something like this in such an old-school language.
A new project on Hacker News is pitching a big promise in a tiny package: an AI helper library called Adam that can be dropped into apps with a single include file and run across phones, desktops, browsers, and even local machines. On paper, it’s a kitchen-sink release — chat, memory, voice, tools, image handling, database tricks, and more — all bundled into one low-level code library.
But the real fireworks weren’t about what Adam can do. They were about why it was built in C at all. That became the instant comment-section battleground. One camp basically said, “Impressive, but are we really writing chatty AI assistants in one of the oldest, fussiest programming languages around?” Critics called the choice awkward and painful for something mostly waiting on outside services. In plain English: people think C is great for speed, but not exactly the language you pick for comfort.
Then came the comedy. One commenter compared the naming choices to someone launching a startup called Google, while another accused the creators of pulling off “advanced SEO” by naming the company sqliteai and the library Adam. Translation: the branding drama may be almost as spicy as the code. Even the praise had a raised eyebrow attached — lots of “seems neat though” energy. So yes, Adam impressed people, but the comments turned it into a classic internet spectacle: part admiration, part confusion, part roast session.
Key Points
- •Adam is introduced as an embeddable AI agent library in C with a unified interface for cloud APIs and local models.
- •The library includes an end-to-end agent loop with tool calling, memory, sessions, voice, streaming, and structured output.
- •It supports multiple providers and runtimes, including Anthropic, OpenAI, Google Gemini, Groq, Together, xAI, and local GGUF models via llama.cpp.
- •The feature set includes local vision, image generation, database extensions for SQLite and PostgreSQL, long-term memory, multi-agent workflows, guardrails, caching, and a filesystem sandbox.
- •The project provides build targets, testing instructions, API documentation, and example programs for chat, tools, local models, vision, voice, and memory workflows.