Workflow & graph engines
graph, DAG and state-machine runtimes for agent workflows
langchain-ai/langgraph LangGraph (langchain-ai/langgraph) is the clear leader for open-source agent workflow engines; if you need durable, stateful, long‑running agent orchestration with human‑in‑the‑loop support and a large ecosystem, start there. Choose alternatives only when you have specific requirements around self‑hosting, vertical integrations, strict auditability, or specialized performance tradeoffs.
Which one matches your setup?
Answer any of the questions — the shortlist updates as you go. Recommendations come from the capability passports below, nothing else.
Comparison matrix
Axes are extracted from each project's docs by our review pipeline; the maturity score is computed from stars, growth and commit activity — not an opinion. Click a column to sort.
| Runs in | Models | Context | Cost to run | |||||
|---|---|---|---|---|---|---|---|---|
⭐ 41.0k +429/7d | CLI | Fixed provider | Diff only | None mentioned | ●●●●● | Third-party cloud | ●●●●● | Free, open-source |
⭐ 28.0k +42/7d | Web appCI | Fixed provider | Flow-level only | Real-time validation | ●●●●● | Self-hostable | ●●●●● | Free, self-hosted |
⭐ 3.2k +114/7d | CLI | Fixed provider | Whole-repo analysis | Severity gating | ●●●●● | Self-hostable | ●●●●● | Free, self-hosted |
⭐ 8.9k | Web app | BYOKFixed provider | Diff only | No explicit controls | ●●●●● | Self-hostable | ●●●●● | Self-hosted (free) with optional Astron Cloud |
⭐ 3.1k +27/7d | Coding-agent pluginCLI | OpenAIAnthropic | Whole-repo analysis | Acceptance gating | ●●●●● | Depends on cloud agents | ●●●●● | Requires cloud agent provider (may incur provider cost) |
⭐ 3.0k +356/7d | CLICoding-agent pluginCI | OpenAIAnthropic | Diff + related files | Approval & strict checks | ●●●●● | Self-hostable | ●●●●● | Free to run locally; external agent API costs may apply |
⭐ 1.8k +76/7d | CLI | OpenAIAnthropicGeminiFixed provider | Diff only | Self-evaluation loops | ●●●●● | Cloud APIs (your key) | ●●●●● | Your API keys for cloud LLMs (Claude/GPT/Gemini); optional local model downloads |
⭐ 1.4k +113/7d | CLICoding-agent pluginWeb app | BYOK | Diff + related files | Evidence gates | ●●●●● | Self-hostable | ●●●●● | Your API key, per-use |
⭐ 195 +2/7d | CLIIDEWeb app | BYOKOpenAIAnthropicGeminiOllama / local | Workspace-level | Human review gating | ●●●●● | Self-hostable | ●●●●● | Free to run locally or via the hosted web app; external model API usage or hosted services may incur provider costs. |
⭐ 1.9k | CLI | BYOKOpenAIFixed provider | Diff + related files | None noted | ●●●●● | Self-hostable | ●●●●● | Your API key / provider billing |
⭐ 639 | CLI | BYOK | Related files | Structured output | ●●●●● | Your API key | ●●●●● | Your API key (OpenAI-compatible endpoint) |
⭐ 619 +2/7d | CLI | BYOKOpenAIAnthropicGeminiOllama / local | Config & strategies | Confidence & risk filters | ●●●●● | Self-hostable (local LLMs) | ●●●●● | Your API keys (OpenAI/Anthropic/Groq/Ollama etc.), pay-per-use cloud fees as applicable |
Capability profiles
Six axes, 0–5 each. The shape tells you the strategy: a wide hexagon is a generalist, a spike is a specialist. Showing the 8 most established — the rest are in the full catalog.
Provides low-level orchestration for long-running, stateful agents with durable execution and human-in-the-loop capabilities that typical stateless workflow engines do not emphasize.
An event-driven, open-source orchestration platform that combines a declarative YAML workflow model with a rich plugin ecosystem and a visual editor for building and running workflows across environments.
Orchestrates and verifies multi-step work by discovering and assigning local Skill modules and enforcing end-to-end checks with resumable state.
Enterprise-grade agentic workflow platform with built-in RPA and high-availability focus for production deployments, unlike many prototyping-focused workflow engines.
Combines a public, curated catalog of reusable agent loops with an installable in-agent skill that discovers, crafts, audits, runs, and prepares loops for publication.
Implements the Depth Tree decomposition with a strict, evidence-backed runnable gate ledger that enforces approvals and re-verification for substantial agent tasks.
Graph-powered, agentic workflows for end-to-end multimodal video understanding, editing and remaking with iterative self-evaluation.
Provides a local-first operating layer for Hermes that enforces explicit evidence boundaries and verification gates while keeping executors and provider access observable.
All repositories (13)
Event Driven Orchestration & Scheduling Platform for Mission Critical Applications
Enterprise-grade, commercial-friendly agentic workflow platform for building next-generation SuperAgents.
Intelligent Skill routing and workflow orchestration for AI agents — +21.12 pp reward, −29.6% tokens on Skills…
A library of practical AI-agent loops and an installable skill for finding, adapting, and designing repeatable…
Anti-laziness skill for AI agents. Core: the Depth Tree method, which splits a task N layers deep and gives ev…
[ICML 2024] LLMCompiler: An LLM Compiler for Parallel Function Calling
[EMNLP2026] "VideoAgent: All-in-One Agentic Framework for Video Understanding and Editing, and Remaking"
Just install once. The coding intelligence, a long-term memory system, optimized workflow packages, only plugi…
An open-source LLM based automatically daily news collecting workflow showcase powered by Agently AI applicati…
AI-Hedge-Fund for Crypto 🚀 AI-powered hedge fund for cryptocurrency trading, leveraging LLM agents for intel…
Server-enforced workflow discipline for AI agents. An MCP server providing persistent work items, dependency g…
AI-driven workbench for computational materials science — interactive 3D structure viewer, natural-language Ca…
Hosted alternatives
If running your own reviewer is more ops than you want, these managed services cover the same job.
LangChain's hosting for LangGraph workflows: the graph runs for you with persistence, streaming and a debugging studio, so you ship the orchestration code without a server and state store.
Try LangGraph Platform →Managed durable execution for long-running, crash-safe workflows: the same retries and state you would run on Temporal, hosted rather than operated by you.
Try Temporal Cloud →The hosted side of Kestra's event-driven orchestration: declarative workflows with scheduling and a visual editor, run for you instead of a self-hosted deployment.
Try Kestra Cloud →