Organization repository on GitHub
Composable Elixir library for LLM interactions built on Req and Finch
- Forks
- 213
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
Jev directory
CommunityJev Repos & SDKs
ReqLLM
ReqLLM is an Apache-2.0 Elixir library built on Req and Finch that standardizes LLM requests and responses across providers, exposing 1,205 models across 21 integrations behind one unified API.
- Stars
- 597
- Language
- Elixir
- License
- APACHE-2.0
- Reading time
- 2 min
- Published by
- Community
- Last verified
- Oct 2026
Works forJev for Tool calling
Used inJev for Software & DevTools
Highlights
- 1,205 models across 21 implemented provider integrations from LLMDB, or 1,218 across 22 namespaces with google_vertex_anthropic.
- 92 non-text operation models across embedding, image generation, text-to-speech, transcription, rerank, and OCR.
- Install with mix igniter.install req_llm, or add {:req_llm, "~> 1.6"} to mix.exs.
- Req handles request/response calls; Finch handles streaming due to Req's known SSE limitations.
- Optional deps only for amazon_bedrock (ex_aws_auth) and google_vertex (goth); local Ollama and LM Studio need no catalog entries.
Quickstart
elixir
def deps do
[
{:req_llm, "~> 1.6"}
]
endWatch out
Reviewed against the truncated README on 2026-10-15. Model counts, provider coverage, and recorded spec figures are the project's own numbers from LLMDB; the README is cut off mid-sentence in the optional-dependencies section, and it does not mention Jev or TypeSafe.
Questions about this entry
Short answers drawn from the linked project or page.
An Elixir package built on Req and Finch that calls LLM APIs and standardizes requests and responses across providers, with a high-level API inspired by the Vercel AI SDK, including generate_text, stream_text, and generate_object.
The recommended path is mix igniter.install req_llm. Manually, add {:req_llm, "~> 1.6"} to your deps in mix.exs and run mix deps.get; provider callbacks translate model metadata, options, bodies, and responses behind the same public API.
The README reports 1,205 models across 21 implemented provider integrations from LLMDB, or 1,218 models across 22 provider namespaces counting google_vertex_anthropic, spanning OpenAI, Anthropic, Gemini, Bedrock, Groq, OpenRouter, and more.
It tracks 92 non-text operation models across embedding, image generation, text-to-speech, transcription, rerank, and OCR APIs, and its fixture suite holds 622 unique recorded model specs as a compatibility ledger for text and multi-modal provider behavior.
Two do, and both are optional: amazon_bedrock uses ex_aws_auth for AWS Signature V4 request signing, and google_vertex uses goth for Google Cloud OAuth via Application Default Credentials. ReqLLM raises a clear error naming the package if one is missing.
Yes. Local inference is supported through Ollama and LM Studio using model identifiers from your own local server, without requiring LLMDB catalog entries, so a fully local setup is possible for supported models.
No. The README covers providers, installation, model catalogs, and operations without mentioning Jev or TypeSafe; the repository entered this catalog through the logicrw community list.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: An Elixir port of the official SDK
More Jev resources like this
A lightweight, self-hosted LLM proxy that gives Hack Club's teen community free chat, embedding, and image models over OpenAI-compatible endpoints, with Hack Club auth, API keys, moderation, spending limits, and analytics.
Jev Repos & SDKs#community#catalog-import#logicrw
Community
hackclub/aiTypeScript
hackclub/aiTypeScriptA small, idiomatic Elixir client for Jev: Jev.ask?/2 returns a real boolean, choose/3 picks among supplied options, score/3 rates ordered levels, and analyze/3 sends several typed questions in one provider request.
Jev Repos & SDKs#community#catalog-import#madewithjev
Community
elue-dev/jev_elixirElixir
elue-dev/jev_elixirElixirA zero-dependency tool selector that narrows a large pool of agent tools to the relevant few before the LLM call, running offline with TurboQuant or on TypeSafe Jev when a TYPESAFE_API_KEY is set.
Jev Repos & SDKs#community#catalog-import#logicrw
Community
hemanth/tool-pruneJavaScript
hemanth/tool-pruneJavaScriptA multi-provider Jev transport layer with fail-closed validation, used by jkudish/jev-browser and jkudish/jev-mcp. One ask() call picks a carrier, sends typed questions and validates answers instead of throwing a verdict.
Jev Repos & SDKs#community#catalog-import#logicrw
Community
jkudish/jev-agent-toolsJavaScript
jkudish/jev-agent-toolsJavaScriptA Go client for the TypeSafe System One API with zero dependencies outside the standard library: typed Noul, Choice and Score questions whose answers are read back through the handles that asked them.
Jev Repos & SDKs#community#catalog-import#logicrw
Community
2389-research/typesafe-goGo
2389-research/typesafe-goGoComposio is an MIT-licensed SDK monorepo giving AI agents 1,000+ pre-authenticated toolkits with per-user sessions, authentication, triggers, and a sandbox, and it ships TypeScript and Python adapters including one for TypeSafe (Jev).
Jev Repos & SDKs#community#catalog-import#cobanov
Community
composiohq/composioTypeScript
composiohq/composioTypeScriptFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.