开源目录
工具生态
探索基于 Jev 构建的社区开源工具。默认按 GitHub Stars、Forks 及名称排序。
Educational CLI (`aihf`) simulating an AI hedge fund: define mandates of LLM investor agents and strategies, run simulated trading cycles over chosen tickers, and backtest equity curves with anonymized prompts that curb model recall—no real trades.
64,72011,390Laya is a non-autoregressive decision engine that answers typed questions (choice, score, noul) over text or JSON states in a single ~33 ms forward pass, trained with RL on strictly proper scoring rules for calibrated confidence, with a router that dispatches to one of three checkpoints across 100+ languages.
35,6293,199Composio is a TypeScript/Python SDK and CLI that equips AI agents with 1000+ pre-authenticated app toolkits via per-user sessions, runtime tool discovery, triggers, hosted MCP endpoints, a sandbox, and provider adapters for OpenAI, Anthropic, LangChain, and other frameworks.
31,1264,956GPT Researcher is an open-source deep research agent that plans questions, crawls web and local sources in parallel, filters context, and generates detailed, cited, unbiased reports exportable to PDF/Word, usable via Python package, server, Claude Skill, or MCP integration.
30,7514,235An interactive CLI that scaffolds a modular, full-stack, typesafe Next.js application using the T3 Stack—optionally integrating tRPC, Tailwind CSS, TypeScript, Prisma or Drizzle, and NextAuth.js via `npm create t3-app@latest` or equivalent yarn, pnpm, or bun commands.
29,2351,555A cross-platform CLI and C library that lets AI agents read desktop apps via the OS accessibility tree as JSON snapshots with element refs, then act on them (click, type, scroll, window, clipboard, notification control) headlessly on macOS, Windows, and Linux—no pixel matching.
24,8421,853jev-ultrafast is a browser-use tool built for maximum speed, providing ultra-fast browser operations when execution latency is the priority.
24,5611,889DeepEval is an open-source, Pytest-style framework for unit-testing LLM apps—agents, RAG pipelines, chatbots—via locally run LLM-as-a-judge metrics (G-Eval, hallucination, tool correctness), synthetic datasets, benchmarks, and CI/CD integration.
19,5252,215QuantDinger is an open-source, self-hosted AI trading platform that turns trading ideas into Python strategies with multi-provider AI research, server-side backtesting, paper and live execution across crypto exchanges and brokers, plus monitoring via web, mobile, API, and MCP interfaces backed by PostgreSQL, Redis, and Celery workers.
13,8162,817kev is a local decision-model server — a LoRA adapter plus pointer readout head on a Qwen base (0.5B–8B) — that answers many typed questions (yes/no, multi-choice, ordinal score) about a document in one decode-free prefill pass, returning calibrated softmax probabilities through a TypeSafe System One–compatible API.
10,032687An Android accessibility-service chat copilot that reads on-screen WeChat/QQ/X/Feishu conversations (via UI nodes or offline OCR), uses a Jev judgment model to assess intent and risk in ~1s, and ranks three generated reply drafts it fills in but never sends—enriched by local knowledge-base, contact, and chat-history context.
8,4301,346A Claude Code plugin that replaces standard compaction summaries with Jev decision-making, scoring every tool call and result in a single fast request so stale items are dropped or truncated while retained content stays verbatim.
7,989588gltfjsx converts GLTF/GLB models into reusable React Three Fiber JSX components, exposing scene-graph nodes and materials for declarative editing while optionally pruning, instancing, type-checking, animating, and compressing assets for faster web delivery.
5,906382SemIf adds semantic if-condition evaluation backed by open models running locally on a home NVIDIA RTX 3090, enabling meaning-based branching without depending on a hosted model service.
4,963356SemIf is an open-model toolkit for runtime-defined semantic decisions that reads typed option probabilities directly from local or browser LLM logits, with shared-state reuse, calibration, and multiple backends for fast auditable agent branching without generated answer parsing.
4,963356RuVector is a Rust-native substrate providing local embeddings, persistent vector search, graph memory, and feedback learning—via npm CLI, MCP server, Python, and Rust crates—so agents can make fast local decisions and recall context across sessions.
4,708631RubyLLM is a Ruby-native AI framework providing one consistent API across 18 providers for chat, agents, tools, structured output, streaming, embeddings, reranking, image/video/audio generation, transcription, OCR, and moderation, with Rails integration and cost tracking.
4,517539An open-source CopilotKit/AG-UI chat interface whose Python agent answers each request with text, native components, or custom interactive UIs—3D explainers, comparison charts, calculators, and maps—streamed into sandboxed iframes, using your own OpenAI and Jev keys.
4,248441Open-source, local-first CLI that captures AI agent session history across Claude Code, Cursor, Codex, and 20+ harnesses into a unified OpenTelemetry event model, turning workflows, corrections, and debugging patterns into reusable memory retrievable via MCP or Agent Skills.
3,555320jevgrep (`jg`) is a Node.js 22+ CLI that lets coding agents ask natural-language questions about a repository and returns relevant files, reading leads, and verbatim source excerpts in one stdout response, using the Jev model via Vercel AI Gateway to judge relevance.
3,241247CLM serves CLM-8B, a contrastively trained System One model (frozen Qwen3-8B encoder + 20M projection heads), via a TypeSafe-compatible API and playground for typed decisions, candidate ranking, and agentic verification with cached embeddings at up to 9× lower latency.
3,208276A Monad trading bot that queries a TypeSafe Jev model each ~300 ms block for a buy/sell call on Kuru's MON-USDC book and posts a post-only limit order one tick inside the touch—earning spread on taker fills—with dry-run mode, sub-block two-RPC-round-trip hot loop, and an SSE dashboard server streaming per-block quotes, fills, and P&L.
3,138601An open-source non-autoregressive decision engine that produces typed, statistically calibrated decisions across 100+ languages in a single forward pass, delivering millisecond-level latency and automatic per-request checkpoint routing for intent classification, routing, and scoring workloads.
3,075342A collection of agent skills for building with TypeSafe's System One API, giving coding agents reusable guidance and workflows that make it easier to integrate and work with the API correctly.
2,905187memsearch is a cross-platform plugin that gives AI coding agents persistent semantic memory — automatically capturing conversations as editable Markdown, indexing them in Milvus with hybrid dense+BM25 retrieval, and distilling repeated workflows into reusable skills.
2,893289NanoJev is a 0.6B-parameter parallel decision model (Qwen3-0.6B backbone plus decision heads) that maps states, questions, and 2–255 candidates to complete probability distributions in a single forward pass—zero output-token decoding—with persistent serving and verified maze/Snake demos.
2,600262Nimble scores text against a flat schema of enum and boolean questions, returning the chosen answer and per-candidate probabilities from a single forward pass without generating reasoning, and provides the contrastive data-curation and LoRA fine-tuning recipe for its 9B model.
2,311174ContextGem is an open-source Python framework that extracts structured, validated data from documents with LLMs and decision models—generating prompts from plain-language descriptions, tracing each fact to its source, and scoring closed questions with answer probabilities.
2,091189@ts-react/form is a ~3kb headless React library that auto-generates fully typesafe forms from zod schemas via a schema-to-component mapping, eliminating react-hook-form boilerplate while preserving per-field prop control and custom component rendering.
2,03542LLPhant is a PHP 8.1+ framework for building generative AI applications and vector database (RAG) workflows, offering a simple API that integrates with Symfony and Laravel and supports LLM providers including OpenAI, Anthropic, Mistral, Ollama, LM Studio, and OpenAI-compatible services like LocalAI.
1,721183
没有找到匹配的工具。