waymark
waymark is a Python library that keeps background jobs and multi-step workflows running through crashes and restarts for developers, using only Postgres.
What it is good for
- Keeping long background jobs running after a server restart
- Retrying failed steps in a multi-step workflow automatically
- Running tasks in parallel and pausing a workflow for days
- Scheduling recurring jobs without extra infrastructure
At a glance
- Who it is for
- Setup difficulty
- Developer ยท Requires programming
Level 5 of 5. You need to write code to use it.
- Cost
- Free and open source
- Type
- Framework or SDK
- Tasks
- Works with
- Postgres, Python, pydantic-ai
How to get started with waymark
- Install the package with
uv add waymark. - Write actions as async functions and a workflow class that orchestrates them.
- Set
WAYMARK_DATABASE_URLto your Postgres database. - Set
WAYMARK_HTTP_ENABLED=trueto turn on the built-in webapp. - Start the node with
uv run waymark-executor, then run the workflow from Python.
From the project's own documentation. If a step does not match what you see, the official pages have the latest instructions.
Before you try it
- You need to write Python code to define and run workflows.
- It requires a Postgres database to store workflow state.
- Python is the only supported language for now.
- Workflow code must use supported patterns or it fails to compile.
Momentum
- Rank this week
- Not ranked (last #11, week 45, 2026)
- Best rank
- #11 week 45, 2026
- GitHub stars
- 162
GitHub numbers as of Nov 16, 2026. Updated weekly.
On Hacker News
- Show HN: Waymark, A workflow engine that compiles jobs instead of replaying 88 points, 48 comments, Nov 5, 2026
Ranking history
Official links
Questions
- Is waymark free?
- Yes. It is free and open source (MIT license).
- Do I need to know how to code to use waymark?
- Yes. You need to write code to use it.
- What does waymark work with?
- According to its own documentation, it works with Postgres, Python, and pydantic-ai.
- How do I get started with waymark?
- 1. Install the package with uv add waymark. 2. Write actions as async functions and a workflow class that orchestrates them. 3. Set WAYMARK_DATABASE_URL to your Postgres database. 4. Set WAYMARK_HTTP_ENABLED=true to turn on the built-in webapp. 5. Start the node with uv run waymark-executor, then run the workflow from Python.
- What should I know before trying waymark?
- You need to write Python code to define and run workflows. It requires a Postgres database to store workflow state. Python is the only supported language for now. Workflow code must use supported patterns or it fails to compile.
- What is waymark good for?
- Keeping long background jobs running after a server restart, retrying failed steps in a multi-step workflow automatically, running tasks in parallel and pausing a workflow for days, and scheduling recurring jobs without extra infrastructure.
Similar agents to waymark
- Conductor
A workflow engine for multi-step jobs and AI agents, aimed at the same crash-safe execution goal.
- Agno
Focuses on building and hosting AI agents as services rather than general durable background jobs.
- crewAI
Python framework for teams of AI agents, not for durable background job execution.
- LangChain
Builds LLM-powered apps and agents rather than durable job execution on Postgres.
- Mastra
TypeScript framework for AI agents and apps, not a Python durable job runner.
- litellm
litellm is an open source AI gateway that lets developers call 100+ AI model providers through one OpenAI-style interface with cost tracking and load balancing.