Cloudflare introduces clairic: a category-free model paradigm for extracting values from raw inputs
ritakozlov · x · 2026-10-20
Cloudflare's michellechen unveils clairic, billed as a brand-new model paradigm: a system one model that extracts, computes, estimates, and predicts values directly from unstructured inputs without predefining categories.
Unlike decision models like Clef — where you list options upfront and get probabilities back — clairic assigns field values directly (e.g., outputting "llama" with 0.99 confidence where a classifier could only pick from pig/cow/horse).
The proposed pipeline: clairic extracts values → Clef makes decisions with them → LLMs act on those decisions, all running on Workers AI. Suited for data you've never seen before.
Related event: Cloudflare Launches Clairic, a System One Extraction Model on Workers AI(12 posts)→
More from coding & agent
- Gensyn CEO: auditable training plus inference logs make AI-auditing-AI trustworthy — benfielding · 2026-10-20
- Dev builds a portable memory layer so every personal AI agent knows everything about him — blaizedsouza · 2026-10-20
- Graphify maps your whole codebase into a queryable knowledge graph, hits GitHub Trending #6 — blaizedsouza · 2026-10-20
- Builder claims Opus 5.5 can replace After Effects, publishes motion design studio guide — blaizedsouza · 2026-10-20
- Seroter's reading list: Meta's thin orchestrator agent architecture, agentic verification, and why personal software factories waste time — rseroter · 2026-10-20
- Maestro leaderboard: agent hits month-long unattended runtime in 11.8 days — Dave_Maynor · 2026-10-20