# Cloudflare previews Clairic: typed extraction and forecasting without generative output tokens Cloudflare's October 19 Clairic preview adds a non-generative extraction model to Workers AI, returning typed spans, dates, numbers and forecast intervals in a single pass. It promises lower-cost document processing, but Cloudflare warns that complex arithmetic and forecasting remain unreliable. Clairic joins Clef as a specialised one-pass model on Workers AI; typed extraction has a different failure model from LLM-generated JSON, with vendor-only benchmarks and preview limitations. - Status: Active - Published: 2026-10-20T17:36:21+13:00 - Updated: 2026-10-20T17:36:21+13:00 - Categories: Artificial Intelligence, Web Development, Cloud & Infrastructure, AI Models, Open Models, Inference & APIs, Developer Tools - Tags: AI models, API pricing, Cloudflare, Workers AI - Canonical HTML: https://beyondthe.news/dossiers/cloudflare-clairic-workers-ai-system-one-extraction-preview ## What changed On October 19, 2026, Cloudflare announced Clairic, the first hosted preview of its System One extraction-model class, available through Workers AI as @cf/cloudflare/clairic. Instead of generating prose tokens, it answers named fields with typed text spans, numbers, dates, durations, lists, estimates or forecasts. Outputs include confidence, p_null and probability intervals for some numeric tasks. Cloudflare intends to release open weights later, but the preview is currently hosted and the weights are not yet available. ## Why it matters A common application pattern asks a general LLM to extract invoice fields, parse dates, classify a claim, then format JSON. Clairic attempts to make the extraction portion a cheap typed operation with spans tied to input and predictable output shapes. That could simplify ingestion pipelines and reduce latency, but type correctness does not mean factual correctness; operators still need validation, confidence thresholds and deterministic arithmetic for important decisions. ## A schema-bound extraction API A request sends state text and up to four optional images, plus a map of fields. Types include text spans with character offsets, number, date, time, datetime, duration, list, bounded estimate and multi-step forecast. The API returns declared types or null rather than free-form generated text. It can split large field maps into parallel passes that share the input. ## One-pass economics and reported tests Cloudflare says Clairic is billed for input processing without generated output tokens, and most calls take less than a second. In its own 21-day TSA passenger-count experiment, mean absolute percentage error was about 3.3%, with 17/21 observations inside its nominal 80% interval. In a staged customer-support pipeline combining Clairic, Clef and GLM-5.3, Cloudflare reports 19.3 seconds and $6,032 per million claims versus 61.7 seconds and $11,740 for an all-LLM pipeline. These are vendor-run demonstrations, not independent benchmarks or guaranteed savings. ## Where not to trust the preview Cloudflare says one- and two-step calculations work better than long arithmetic chains; estimates require contextual evidence, interval calibration can drift on unfamiliar data, and forecasting is not a replacement for a dedicated time-series model. A syntactically valid answer may still be wrong. Use p_null, confidence and intervals to route uncertain outputs to review, and use application code for deterministic financial arithmetic. ## Deployment and open-weight boundaries The currently documented route is a hosted Workers AI preview at @cf/cloudflare/clairic via Worker env.AI.run or REST. Open-weight release is an intention, not a shipped artefact. Test latency, cost and accuracy against your own documents before switching from general LLM extraction. ## Key details - Announced October 19, 2026. - Workers AI model identifier: @cf/cloudflare/clairic, hosted preview. - Typed outputs: text spans with offsets, numbers, dates, durations, lists, estimates and forecasts. - Outputs include confidence and null probabilities; forecast/estimate intervals are available. - Cloudflare reports 3.3% mean error and 17/21 interval coverage on one TSA forecast demonstration. - Cloudflare reports a staged pipeline at 19.3s versus 61.7s and $6,032 versus $11,740 per million claims, not independently verified. - Open weights promised for later, not currently released. ## Builder takeaways - Use a specialised extraction stage for predictable typed fields rather than requiring an LLM to produce arbitrary JSON. - Preserve source offsets and confidence scores for human review and audit. - Do arithmetic and financial validation in code; do not rely on a preview model for chained calculations. - Benchmark on your own documents and failure modes; vendor demo figures are not production guarantees. - Do not label Clairic open-weight until Cloudflare actually publishes weights. ## What to watch - Public release of Clairic weights, licensing and reproducible evaluation assets. - Published hosted pricing and production SLAs after preview. - Independent accuracy, interval calibration and document-extraction comparisons. - Improvements to chained arithmetic and forecasting. ## Uncertainties - Only vendor-provided demonstrations and benchmark figures are currently available. - Weights, licence and self-hosting instructions have not shipped. - Preview pricing, availability and performance may change. - Forecast interval coverage may degrade out of distribution. ## Sources - [Introducing Clairic, a new System One class of extraction models](https://blog.cloudflare.com/clairic-extraction-model/) — Cloudflare Engineering · primary/vendor. October 19 launch, API, examples, evaluation and explicit preview caveats. - [Introducing Clef: our open-source decision models](https://blog.cloudflare.com/introducing-clef-decision-models/) — Cloudflare Engineering · primary/vendor. Context for distinct decision-model category; Clairic is extraction rather than classification.