User repository on GitHub
SillyTavern extension that measures each reply and instructs the narrator only when a rule matches.
- Forks
- 6
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Integrations & Adapters
ST-jeved
SillyTavern extension that sends chat text to Jev sensors and fires one rule action per turn such as nudging, rerolling, editing lists, or running scripts.
- Stars
- 33
- Language
- JavaScript
- License
- MIT
- Published by
- Community
- Author
- mossyfield
- Last verified
- Oct 2026
About this project
Jeved is a SillyTavern extension. It sends chat messages to Jev, a small decision model, with questions that you write. A rule reads the answers and acts: it adds one instruction to the prompt, rerolls the reply, changes a list, or runs an STscript. When no rule matches, Jeved adds nothing. Pacing: ask for an event when the story stays calm. Steering: read the message you send and shape the reply to it. Rerolls: reroll a reply that speaks for your character or breaks a rule that you wrote.
Quoted from the project's README. Read the full README.
Used inJev for Software & DevTools
- Languages
- JavaScript
Highlights
- Score, Choice and Noul sensors read latest turn, player message, history and picked context.
- Rules fire on N of last M messages with thresholds, exceptions, cooldowns and priority order.
- Actions nudge prompts, reroll once per turn, edit per-chat lists, or run STscripts.
- Lists, activity chart and cost view track entries, last replies and shared calls.
Watch out
Needs SillyTavern 1.18 or later plus a configured host key, and reroll rules do not run in group chats. Copy reviewed against the repository README on 2026-10-04, not from running it.
Questions about this entry
Short answers drawn from the linked project or page.
No. A failed call never stops chat: pre-reply measurement waits at most 2 seconds and post-reply sensing caps at 3 seconds, after which the reply goes out with no instruction.
Sensors are explicit Score, Choice or Noul questions over recent messages plus picked context, and rules test them with thresholds, exceptions and cooldowns. Preview applies drafts to stored answers with no API call.
The messages and context the sensor picked plus the batched sensor questions. Sensors sharing counts and context share one call, and macros like user names resolve before sending.
Order is priority and only the first matching reroll rule fires once per turn, with cooldowns holding individual rules. Nudge edits only the prompt copy, not the chat message itself.
Three types. Score takes a scale of two to ten steps and returns a number from zero to the last step, such as 1.4; Choice takes two to 255 options and returns one option name; Noul takes a statement and returns the chance from 0 to 100 percent that it is true. A sensor is measured when an enabled rule uses it, or when you tick Measure anyway.
Pick a list under Repeat over list and write the entry placeholder in the question, scale, or options, and Jev answers once per entry inside the call the sensor already makes. A rule on it tests each entry and fires once for the turn, and the matching entries reach the instruction, script, and list value as entry, entries, and entries_json.
Jev takes about 32,000 tokens per call, and the Context cap limits what a sensor sends. Over the cap Jeved drops, in order: world info, chat examples, preset prompts last first, persona, character note, scenario, then personality, and finally shortens the description. It never cuts the main prompt or the post-history prompt.
More Jev resources like this
25GitHub stars
Agent runtime with a terminal UI and a messaging gateway that keeps memory, self-built skills, session search, scheduling, and subagents across Telegram, Discord, Slack, WhatsApp, Signal, and CLI.
Jev Integrations & Adapters#community#jevbest#catalog-import
Community
wundercorp/lokiPython
wundercorp/lokiPython14GitHub stars
Python library for schema-enforced JSON extraction into Pydantic models, with provider-native JSON mode fallback, auto-repair, token and cost tracking, plus agents, RAG, and eval helpers.
Jev Integrations & Adapters#community#jevbest#catalog-import
Community
jhd3197/PrompturePython
jhd3197/PrompturePython12GitHub stars
Programming language and VM where agents run as isolated actors with managed context, durable effects, and confidence-aware control flow, including a Jev decision effect.
Jev Integrations & Adapters#community#jevbest#catalog-import
Community
ekizito96/TurnRust
ekizito96/TurnRust19GitHub stars
Ruby gem that models Jev as field predicates: feels and decide bind to model attributes with shared definitions, thresholds, confidence gates, batching, and test tapes.
Jev Integrations & Adapters#community#jevbest#catalog-import
Community
Qew7/jev-feelsRuby
Qew7/jev-feelsRuby8GitHub stars
Shadcn registry search block that shows lexical hits instantly then re-ranks the top candidates with Jev intent scoring in one request, with keyword fallback on failure.
Jev Integrations & Adapters#community#jevbest#catalog-import
Community
kylemclaren/jevsearchTypeScript
kylemclaren/jevsearchTypeScript4GitHub stars
DuckDB table functions over a VGI worker that expose Jev Choice, Noul, Score, and multi-question ask for LATERAL joins with confidence-based routing.
Jev Integrations & Adapters#community#jevbest#catalog-import
Community
Query-farm/vgi-typesafePython
Query-farm/vgi-typesafePythonFrom the Jev community
Posts from builders shipping with Jev right now.