User repository on GitHub
Use LLM judgment in regular Elixir control flow.
- Forks
- 1
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
CommunityJev Integrations & Adapters
Gut — vinibrsl
Elixir DSL where Gut.feel/4 maps a subject and question to one offered choice, using Jev, any ReqLLM model, local Ollama, or a custom adapter.
- Stars
- 18
- Language
- Elixir
- License
- APACHE-2.0
- Published by
- Community
- Author
- vinibrsl
- Last verified
- Oct 2026
About this project
Gut is a simple DSL for LLM decisions in Elixir. It picks an Elixir value for your subject and question. No agent framework required, just a few lines of code and your LLM of choice. Add gut and reqllm to mix.exs: Choose one of these options to configure the adapter: Configure Gut.ReqLLM.Jev to use a Jev evaluation model: Use Gut.ReqLLM with any ReqLLM-supported model: With Gut.ReqLLM, options other than :model pass to ReqLLM.generatetext/3. Start Ollama and pull a model:
Quoted from the project's README. Read the full README.
Used inJev for Software & DevTools
- Languages
- Elixir
Highlights
- Keyword choices carry descriptions, plain lists need none, integer ranges score 1..5, and allow_unsure routes to review.
- Backends include Gut.ReqLLM.Jev, any ReqLLM model, Ollama at localhost:11434, or a module implementing Gut.Adapter.
- Structured errors use timeout, rate_limited, unauthorized, invalid_answer, and adapter_error with a cause for logging.
- Gut.Test stubs per test process select another choice, and telemetry events omit subjects, questions, and errors.
Watch out
Structs must derive Gut.Subject to limit sent fields, stubs belong to the test process, and Elixir 1.18 or later is required. 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.
Gut narrows the call to one offered choice with typed returns, subject encoding, per-process test stubs, and telemetry, while extra ReqLLM options still pass through to generate_text.
It returns {:error, %Gut.Error{}} with a reason for control flow and the original error in cause for logging, rather than raising or returning an unlisted value.
No. Stubs belong to the test process, tasks it starts can use them, other processes need their own stub, and tests can use async: true.
No. Events carry adapter, model, duration, outcome, and kind, while the README states Gut emits no subjects, questions, choices, errors, or stack traces.
Add gut and req_llm to mix.exs, then configure an adapter: Gut.ReqLLM.Jev for a Jev model, Gut.ReqLLM for any ReqLLM model, Gut.ReqLLM with an ollama: model for local Ollama at localhost:11434 with no key, or a module implementing Gut.Adapter.
Pass an integer range such as 1..5 for a score, and add allow_unsure: true to a list or keyword call so a third result, :unsure, routes the item to review instead of forcing a choice. The returned value is one of the choices you offered.
Derive Gut.Subject on a struct with only: [:subject, :body]. Strings are sent as plain text and other values are encoded as JSON, so deriving the subject is how you limit which fields of a struct reach the model.
Other listings of this project
Forks and copies of this project in the directory. This entry is the canonical one.
6GitHub stars
Python helper that turns judgment calls into one-liners like likely, classify and rate, defaulting to Jev while also running on local or OpenAI-compatible backends.
Jev Models & Reimplementations#community#github
Community
Kungie/gutPython
Kungie/gutPythonMore Jev resources like this
38GitHub stars
File-first project notebook where many agents share context in .context/ files while Jev only judges the next step and records it as JSON.
Jev Integrations & Adapters#community#jevbest#catalog-import
Community
pedrocivita/tocketTypeScript
pedrocivita/tocketTypeScript34GitHub stars
Self-hosted agent runtime that adds an optional Jev judgment layer to memory filtering, lesson dismissal, request classification, and skill selection, with inspectable context and execution.
Jev Integrations & Adapters#community#jevbest#catalog-import
Community
matrixorigin/AstraRust
matrixorigin/AstraRust12GitHub 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/TurnRust3GitHub stars
Task plugin that puts TypeSafe's System One endpoint behind a QuantumNous/new-api channel, so clients keep posting `/v1/systemone` while new-api handles auth, routing, model mapping and token billing against either TypeSafe or the Vercel AI Gateway.
Jev Integrations & Adapters#community#jevbest#catalog-import
6GitHub stars
MCP layer that lets an agent harness ask Jev for a bounded choice from a candidate set it already allows, while permissions and execution stay with the host.
Jev Integrations & Adapters#community#jevbest#catalog-import
Community
typakon4/jev-layerJavaScript
typakon4/jev-layerJavaScript2GitHub stars
A Next.js workbench that converts OpenAI-compatible Chat Completions requests into Jev state and typed questions, compares chat output against structured judgments, and publishes saved question packs as proxy routes. The interface is mainly in Chinese.
Jev Integrations & Adapters#community#jevbest#catalog-import
Community
Chandler-Sun/chat2jevTypeScript
Chandler-Sun/chat2jevTypeScriptFrom the Jev community
Posts from builders shipping with Jev right now.