AI for devs, AI for everyone
Something new in AI ships every day — a model, a tool, a skill. Nobody can keep up alone. Neuricon tracks it all: a verified directory of tools, skills, and models, a daily news wire, and a wiki that explains what it all means.
Start here
Directory
Multi-capability AI products — pipelines, platforms, automation — researched, with setup guides, agent compatibility, and a 1–10 trust score that shows its reasoning.
4341 tools verified
Directory
Agent skill packages — .md-file skills for Claude Code and other agents — researched and verified, with setup guides and trust scores.
313 skills verified
Directory
New releases, trending, and popular models across every provider — pricing, context windows, benchmarks, and open vs closed weights.
1948 models tracked
Evergreen
Long-form explainers on the ideas behind the news — how models work, what agents actually do, which evals matter — written in plain language, not marketing copy.
13 deep guides
Daily
The day's AI releases, launches, and papers — model drops, tool launches, research — with source links and zero hype.
Daily news wire
Live
A place to talk tools, share setups, and get help from people tracking the same firehose. The board is open — a Discord server is coming soon too.
Live discussion board
Soon
Neuricon's own AI agent. Ask it anything — find the right model, tool, or skill, or get a plain-language explainer. In the works.
Soon AI guide
The wire
The wiki
Why the best models of 2026 are sparse: how MoE routes tokens to specialists, what it costs you at inference time, and when dense still wins.
Oct 8, 2026 · 8 min read
Anatomy of a working agent — planner, tools, memory — plus the failure modes nobody warns you about and the evals that actually catch them.
Oct 5, 2026 · 11 min read
A look inside our own pipeline: where releases are spotted, how duplicates are merged, and how a changelog becomes a searchable record.
Oct 1, 2026 · 6 min read
The archive
Meridian-2 tops SWE-bench Verified with a 71% agentic solve rate
Meridian AI · Oct 11, 2026
Northloop ships one-command eval harnesses for LLM apps
Northloop · Oct 10, 2026
Desktop agents get their first safety benchmark
The Agent Index · Oct 9, 2026
Ferrous-7B: a mixture-of-experts model that runs on a laptop
Ferrous · Oct 8, 2026
Latent adds trace-level prompt debugging to its observability suite
Latent · Oct 7, 2026
The community
Post the tools, skills, articles, and research papers worth knowing about. Every approved post, every day you show up, every page you explore earns XP — climb from Explorer to Luminary and unlock badges, profile frames, and faster post approval.
Keep us growing
Neuricon is free and independent. Your donation keeps the servers running, the research going, and the directory growing — for everyone.
Payment gateway is being set up. Donations open shortly.
Why Neuricon
01
Every release we track becomes a permanent, tagged, searchable record — linked to its source, its changelog, and the models it competes with. Six months from now, when you need that 7B MoE from October, you'll find it in seconds.
02
Our pipeline drafts every summary, but nothing publishes without human review — and every claim links back to a primary source. When a benchmark number looks too good, you can check it yourself in one click.
100% of items source-linked
03
Keyboard-first search, RSS for everything, and a public API for the whole archive. Neuricon fits the way developers already work — fast, scriptable, no fluff.
Public API
Every tool, model, and news item — searchable, filterable, machine-readable. No key needed to start; CORS-open and rate-limited.
{
"ok": true,
"items": [
{
"name": "CrewAI",
"slug": "crewai",
"provider": "CrewAI Inc.",
"tagline": "Role-based multi-agent
framework — crews of
collaborating AI agents.",
"trust": { "score": 8 }
},
…
]
}/api/v1/catalogTools & skills — filter by q, provider, kind
/api/v1/modelsModels — filter by q, provider, open_weights
/api/v1/newsLatest AI news, fresh daily
/api/v1/export/catalog.jsonFull dataset export
Pollers watch model hubs, changelogs, and registries around the clock.
Entries are researched, scored 1–10 on trust, and human-reviewed before they publish.
Everything lands in a tagged archive you can query forever.
Early access
One email each morning: the releases that matter, why they matter, and links to dig deeper. No spam, unsubscribe anytime.
Oct 16, 2026
Pipeline task: translation. Tags: llama.cpp, gguf, translation, index, base_model:IndexTeam/Index-Echo-S2ST-9B, base_model:quantized:IndexTeam/Index-Echo-S2ST-9B, license:apache-2.0, endpoints_compatible, region:us, conversational. Downloads: 815; likes: 0.
Source: Hugging Face
Oct 16, 2026
Source: Hugging Face
Oct 16, 2026
Source: Hugging Face Blog
Oct 16, 2026
Source: Hacker News