Jev in production › SDKs and workflow integrations

JarvisCore (Prescott Data)

Python multi-agent framework that injects a Jev client as self.decisions for Choice, Score and Noul judgments, with opt-in Jev subagent routing, model-tier selection and RAG passage filtering. Three questions in one call: 434 ms median warm (company).

Open on GitHub ↗

434 ms median warmmeasured, as published by the source
Use
SDKs and workflow integrations
Industry
AI infrastructure
Form
Open-source tool
Stage
In production
Listed
2026-09-20
Found via
github
Repository
Prescott-Data/jarviscore-framework
Stars
192
Forks
12
Last push
2026-10-11
Language
Python
License
Apache-2.0
JarvisCore (Prescott Data) screenshot
jarviscore/docs/assets/demo.gif in the Prescott-Data/jarviscore-framework README, Apache-2.0; shown from GitHub.

The README opens with

The production runtime for multi-agent systems — a peer-to-peer mesh with no central orchestrator, zero-trust credentials, durable state that survives kill -9 , and a built-in catalog of typed integration atoms. Observability included, not upsold.

4 agents, 4 separate processes, no MCP, no orchestrator. They discover each other over SWIM gossip, claim steps from a shared ledger, research the live web, and deliver the brief to Slack — the token never visible to agent code, resolved by the encrypted vault at the call boundary. Real run, time compressed — full 4-minute narrated video · reproduce it with examples/demosynthesizer.py + demonode1/2/3.py

Badge

For the project's own README, linking back here:

Listed in Jev in production

Also used for sdks and workflow integrations