AgentRun
Open-source DSL that freezes repeated agent work into typed, inspectable steps
Emerged finds new products worth a look, daily — what’s new today
Reef joins inference, feedback, training and versioning in one loop, so an agent can get better from real use without a bespoke MLOps stack.
Reef is open-source infrastructure for continually self-improving agents. It connects serving, feedback collection, training and version management, and supports two routes to improvement: updating model weights or evolving the agent harness itself. It installs from PyPI as reef-infra on Python 3.10+, is Apache-2.0 licensed, and is built by the Human Agent Society group with 30-plus contributors.
More like this
DeepSeek Harness
DeepSeek's open-source agent harness where every capability is a plugin
AutoResearch
An open pipeline from ML research idea to reviewed, traceable evidence
no_human
Open-source agent that takes a ticket to a reviewed pull request, with proof
Sphinx
A pre-push gate that asks what the agent's change does before it ships
OpenBot
Self-hosted AI coworkers, each with its own computer and a policy gateway