CommunityJev Integrations & Adapters
Databricks: query a System One model
Databricks docs for the System One route on Unity Gateway. The example model is system.ai.openjev-qwen35-4b, with noul, choice, and score questions.
- Format
- Docs
- Published by
- Community
- Author
- Databricks
- Last verified
- Oct 2026
Used inJev for Data & Analytics
Highlights
- Requires a Unity Catalog workspace, Unity Gateway, and permission to run the model service.
- The route does not accept a plain model-provider service or a serving endpoint outside Unity Catalog.
- The sample request sets model to system.ai.openjev-qwen35-4b and sends noul, choice, and score questions.
- The docs say the returned model id can differ from the fully qualified service name, and that numbers vary by request.
Watch out
The worked example is an open Jev-like service on Databricks, not a call to api.typesafe.ai.
Questions about this entry
Short answers drawn from the linked project or page.
No. The entry copy states the route does not accept a plain model-provider service or a serving endpoint outside Unity Catalog. The worked example is an open Jev-like service hosted on Databricks, not a call to the vendor's own endpoint.
A Unity Catalog workspace, Unity Gateway, and permission to run the model service. That is a heavier setup than an API key, so read it as a deployment path for platform teams rather than a quick local experiment.
The sample sets the model to an open Jev-like service in the platform's system namespace and sends noul, choice and score questions in one request. So all three primitives go out together, which is the shape to expect from a single pass.
Not literally. The entry copy says the docs warn the returned model id can differ from the fully qualified service name, and that numbers vary by request. Treat the id as a label and any single result as one sample, not a reproducible measurement.
A curl POST to https://<workspace_host>/ai-gateway/typesafe/v1/systemone authenticated with -u token:$DATABRICKS_TOKEN, with a JSON body of model, state and questions. The worked example names system.ai.openjev-qwen35-4b, passes a state object with message and channel fields, and sends one noul, one choice and one score in the same call.
A noul returns a number from 0 to 1, the probability the proposition is true. A choice returns the selected option, a probability for every option and a confidence, with 1 to 255 options defined. A score returns a probability-weighted score, a legend mapping level indexes to your criteria, per-level probabilities and a confidence, with 1 to 10 levels. The response adds a model field and usage token counts.
The route uses a Databricks token, passed as token:$DATABRICKS_TOKEN in the caller's user field of the curl invocation. Request validation errors return HTTP 422. The docs also state plainly that response time depends on the model service and request load, so there is no fixed latency figure to design against, and they point at the separate ai_decide function for decisions over table rows in SQL.
More Jev resources like this
Databricks' 30 September 2026 announcement of ai_decide, a beta SQL and REST function for typed decisions, described as compatible with the TypeSafe API.
Jev Integrations & AdaptersArticle#community#article#databricks
Community
4GitHub stars
Loadable SQLite C extension and Python package that exposes Jev Noul, Choice, and Score judgments as SQL scalar functions and a jev_rows virtual table which batches up to 40 rows into one shared state and joins typed results back to the source by rowid.
Jev Integrations & Adapters#community#sqlite#sql
Community
mgaitan/sqlite-jevC
mgaitan/sqlite-jevC88.2kGitHub stars
Real-time geopolitical intelligence dashboard that uses TypeSafe Jev questions to score news headline severity into five threat levels and categorize events across 14 conflict, cyber and infrastructure domains, alongside news aggregation and a Country Instability Index.
Jev Integrations & Adapters#community#osint#geopolitics
Community
koala73/worldmonitorTypeScript
koala73/worldmonitorTypeScript33kGitHub stars
Alibaba's multi-agent platform, which implements native TypeSafe Jev classification through JevClassifierModel for binary, choice and score routing across agent pipelines. Python 3.11+, Apache-2.0, with a FastAPI agent service and web UI.
Jev Integrations & Adapters#community#multi-agent#framework
Community
agentscope-ai/agentscopePython
agentscope-ai/agentscopePython689GitHub stars
The Vercel AI SDK provider for OpenRouter, including openrouter.decisionModel for typesafe/jev-1.13 evaluations with providerMetadata answers.
Jev Integrations & Adapters#community#typescript#openrouter
12GitHub stars
CLI-installed Gmail classifier that runs 24/7 on user-owned Google Apps Script: the worker asks Jev for three typed judgments per thread (requires_action, is_important, bucket), applies labels, stars, and optional archiving, and routes ambiguous or low-confidence mail to a Review label.
Jev Integrations & Adapters#community#gmail#email
Community
vynnlee/jev-mailTypeScript
vynnlee/jev-mailTypeScriptFrom the Jev community
Posts from builders shipping with Jev right now.