Cette page est affichée en anglais. Une traduction française est en cours.
WorkflowsMay 13, 2026·3 min de lecture

GraphBit — Type-Safe Agentic Framework (Rust + Python)

GraphBit is an open-source agentic framework for deterministic, concurrent execution with type safety, guardrails, observability, and multi-LLM support.

Agent Toolkit
Agent Toolkit · Community
Prêt pour agents

Cet actif peut être lu et installé directement par les agents

TokRepo expose une commande CLI universelle, un contrat d'installation, le metadata JSON, un plan selon l'adaptateur et le contenu raw pour aider les agents à juger l'adaptation, le risque et les prochaines actions.

Native · 94/100Policy : autoriser
Surface agent
Tout agent MCP/CLI
Type
Workflow
Installation
Pip
Confiance
Confiance : Established
Point d'entrée
pip install graphbit
Commande CLI universelle
npx tokrepo install ab5e8e95-b35a-5d16-9ead-43accef894d8
contrat d'installationJSON metadataplan adaptateurcontenu raw
Introduction

GraphBit is an open-source agentic framework focused on deterministic, concurrent execution with guardrails and observability. It provides a pip install graphbit quick start, and the repo is GitHub-verified at 537★.

Best for: building agent workflows that need determinism, concurrency control, and production-grade guardrails

Works with: Python package + Rust runtime; multi-LLM providers; tool functions with descriptions; workflow graphs

Setup time: 10–20 minutes

Key facts (verified)

  • README includes a benchmark table claiming large CPU/memory multipliers vs other frameworks (as stated in README).
  • README lists supported Python versions 3.9–3.13 and shows pip install graphbit.
  • README includes a workflow example with tool decorators and node graphs.
  • GitHub: 537 stars · 115 forks; pushed 2026-04-28 (GitHub API verified).

Main

When evaluating an agent framework, test three things early:

  • Can you reproduce outputs (determinism) under retries?
  • Can you cap concurrency and costs under load?
  • Do guardrails and observability give you enough signal to debug failures?

Use GraphBit's quick start to build a small workflow graph and add one real tool call (filesystem/API) before committing to a migration.

README excerpt (verbatim)

GraphBit - High Performance Agentic Framework

Logo

GraphBit - Developer-first, enterprise-grade LLM framework. | Product Hunt GraphBit - Developer-first, enterprise-grade LLM framework. | Product Hunt

Website | Docs | Discord

InfinitiBit%2Fgraphbit | Trendshift
PyPI Downloads

PyPI Build Status PRs Welcome

FAQ

Q: How do I install GraphBit? A: Install with pip install graphbit per README.

Q: What makes it different? A: README emphasizes deterministic, concurrent execution plus guardrails and observability.

Q: Does it support multiple LLMs? A: Yes—README lists multiple providers and shows provider configs.

🙏

Source et remerciements

Source: https://github.com/InfinitiBit/graphbit > License: Apache-2.0 > GitHub stars: 537 · forks: 115

Fil de discussion

Connectez-vous pour rejoindre la discussion.
Aucun commentaire pour l'instant. Soyez le premier à partager votre avis.

Actifs similaires

Koog — Kotlin/JVM Agent Framework with MCP Tools

A JVM agent framework for predictable, fault-tolerant agents, with MCP tool integration; verified 4187★, pushed 2026-05-14.

Workflows
Agent Toolkit

Langfuse Python SDK — Trace LLM Apps

Langfuse Python SDK adds tracing and observability to any LLM app via decorators or low-level calls, so you can track latency, cost, and prompts.

Workflows
Agent Toolkit

1Code — Multi-Agent Desktop for Claude + Codex

1Code is an open-source coding agent client with worktree isolation, diff previews, model/provider BYOK, and MCP/plugin management (5,522★).

Workflows
Agent Toolkit

RagaAI Catalyst — LLM Eval + Tracing SDK

RagaAI Catalyst is a Python SDK for managing LLM projects with evaluation, dataset management, trace/agentic tracing, and prompt/guardrail workflows.

Workflows
Agent Toolkit
◈Accueil🔍Rechercher👤Moi