How does the node choose a workflow template?

You configure a choice question whose options are your workflow templates, and the node returns the selected option plus its probability distribution, a confidence value and a weighted score that downstream nodes can branch on, replacing keyword matching with a model decision.

What question types can the ComfyUI node ask?

Noul returns a yes/no probability, Choice picks one of up to 255 options, and Score rates content on an ordered 2-10 level rubric. All questions in one request are evaluated in parallel, and each answer comes back with a calibrated confidence value.

How do I install the node, and where does it get its key?

Copy typesafe_jev_node.py into /custom_nodes/, let it pip install typesafe-sdk on first run or install it yourself, restart ComfyUI, then put the key in the api_key parameter or export TYPESAFE_API_KEY. Get the key at console.typesafe.ai.

What do the four outputs look like in practice?

A choice question returns decision_value technical, full_probabilities as a JSON string like {"billing":0.12,"technical":0.85,"sales":0.03}, confidence 0.89 and weighted_score 0.85. A noul returns decision_value noul=0.93 with an empty distribution, and a score returns the chosen level plus its weighted value. question_id is for your code only and is never sent to the model.

How is criteria_json written for each question type?

A noul takes an object with true and false description strings. A choice takes an object of option key to description, up to 255 options, with null for options needing no description. A score takes an array of level labels, minimum 2 and maximum 10, indexed from 0.

How fast is it, what does it cost, and how do rate limits behave?

The README puts typical latency at 70-500 ms per call against 2-10 seconds for LLMs, roughly 193 times faster for decision tasks, at $0.042 per million input tokens with output free. The node retries 429 and 529 with exponential backoff, and confidence summarizes distribution concentration rather than whether your workflow is correct.

Can I use the same decisions outside ComfyUI?

Yes. The same SDK works directly from Python with TypeSafeClient and the Choice, Noul and Score constructors, or from TypeScript with systemOne. The README's own advice is to let Jev handle routing and gating and keep an LLM for generation, and to validate against your own data.

CommunityJev Integrations & Adapters

TypeSafe Jev + ComfyUI

Custom ComfyUI node that routes a prompt to one of three workflow templates using a Jev decision instead of a keyword match. Supports noul, choice (up to 255 options) and score (2-10 levels) questions with parallel evaluation, exponential-backoff retries and full probability distributions as node outputs.

See highlights

Open github.com

Stars: 0
Language: Python
Published by: Community
Last verified: Oct 2026

youneselfakir0/typesafe-jev-comfyui-integration

User repository on GitHub

View on GitHub

TypeSafe Jev + ComfyUI integration with 3 workflow templates

Forks: 0
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

A custom ComfyUI node that interfaces with TypeSafe's Jev System One model, enabling fast, structured decisions from within ComfyUI workflows. noul — Yes/no probability (0 to 1), with confidence score. choice — Pick one option from up to 255 options, with full probability distribution. score — Rate on an ordered rubric (2–10 levels), with probability-weighted value and confidence. Parallel processing — All questions in a single request run in parallel.

Quoted from the project's README. Read the full README.

Tagscommunity comfyui image-generation routing

Works forJev for Routing Jev for Classification Jev for Scoring

Used inJev for Media & Entertainment

Languages: Python
Capabilities: Requires API keySelf-hosted

Highlights

Ships three ComfyUI workflow templates and one Python custom node file.
Choice supports up to 255 options; score supports 2-10 ordered rubric levels.
Outputs decision_value, full probability JSON, confidence and a weighted score.
The node auto-installs typesafe-sdk on first run and retries 429/529 errors with backoff.

Watch out

The GitHub API reports no license for this repository, so reuse terms are unclear. It needs a TypeSafe API key and sends the node's state text to the hosted TypeSafe service; the README itself advises pairing Jev with an LLM for generation.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

Facebook: Intake Router keeps diagnosis out of the answer space at the front desk.

FacebookClinicnotes

YouTube: Jev explained for routing and judging, with a comparison against LLMs.

Hacker News: Vercel on using Jev for defined decisions and GPT-6 Astra when the job also has to write or use tools.

flashbrew4 pointsTypeSafe AI Jev vs. GPT-6 AstraRead the thread on Hacker News

Hacker News: ASTGL replay chose local Laya routing over hosted Jev on one Mac Studio setup.

Jmeg8r15 pointsDoes an open-weight decision model beat a hosted one? Jev vs. LayaRead the thread on Hacker News

Related terms

Glossary definitions related to this entry.

Intent routing

More Jev resources like this

AgentScope

33k

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

AutoGPT

187.5k

Autonomous agent platform with 187,504 stars for building and running agents from plain English or a visual builder. Its platform features first-class TypeSafe Jev decision blocks for typed routing, filtering, scoring and confidence-gated next-action dispatching.

Jev Integrations & Adapters#community#autonomous-agents#platform

Community

Significant-Gravitas/AutoGPTPython

Jev-Mail

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

JIT-JEV Context OS

Epistemic runtime for autonomous coding agents that uses Jev as a sub-100ms System 1 gate to shadow-evaluate capsule facts, propose evictions before bloat and prefetch memory between turns. Its fixture run reports 31.3% fewer turns and 52.6% fewer blind discovery operations, with a fail-open circuit breaker.

Jev Integrations & Adapters#community#agents#context-engineering

Community

wojciechwiesner/jit-contextPython

WorldMonitor

88.2k

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