Our take
Archivedarchived by its owners on GitHub, read-only- Activity
- 84
- Community
- 73
- Issues
- 83
- Pull requests
- 72
-1 stars in the last 30 days
- Releases more often than 84% of the projects we track
- Closes issues faster than 81% of the projects we track
This is a broad, actively maintained capability library, but this repository itself is archived, so new adoption should follow the project into Pydantic AI.
Good fit if
- You already use Pydantic AI and want composable capabilities for coding, research, memory, or context management.
- You want ready-made agent stacks that can also be taken apart and configured as individual capabilities.
- You are comfortable using the maintained version from the Pydantic AI repository, rather than treating this archived repository as the current development home.
Look elsewhere if
- You need an actively maintained standalone GitHub repository to track issues and submit changes to, since this repository is read-only.
- Your agent does not need the extra tools and runtime support beyond Pydantic AI's core agent capabilities.
- You are not building in Python or do not want to use Pydantic AI as the underlying agent framework.
All health signals
| Last commit | 2026-10-02 (16 days ago) |
|---|---|
| Commits, last 90 days | 100+ |
| Releases, last 12 months | 42 (latest v0.36.0, 2026-09-25) |
| Contributors | 33 (top contributor: 28% of commits) |
| Issues closed, last 90 days | 12+ (typically closed in 0 days) |
| Pull requests merged, last 90 days | 39 (typically merged in 0 days) |
| Project age | 7 months |
Checked on 2026-10-19 with the GitHub API.
Project history
Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.
Stars
954Forks
153Change since Nov 12, 2026: 0 stars·0 forks
Overview
Pydantic AI Harness is a Python capability library built on Pydantic AI, the agent framework from Pydantic. It adds the surrounding tools and runtime support that complex, long-running agent tasks can need, such as access to files and shell commands, planning, memory, delegation, and context management.
The project is intended for developers building Pydantic AI agents that need to do more than call a model and return structured output. Capabilities can be added individually or combined into ready-made agents such as Coder and Researcher. This GitHub repository is archived; the project now lives in the Pydantic AI repository, where it is released alongside Pydantic AI.
Key Features
- Compose self-contained capabilities with Pydantic AI agents, or start with the Coder and Researcher harnesses.
- Give coding agents file operations, shell access, and task delegation.
- Add web search, page fetching, browser automation, and research delegation.
- Provide cross-session memory, searchable conversation history, skills, and repository context.
- Manage long-running context with compaction strategies, tool-output limits, and planning reminders.
- Add safeguards such as input and output guardrails, prompt-injection checks, and spending limits.
- Support persistent and durable runs through integrations and step-persistence backends.
- Trace model and tool activity through OpenTelemetry instrumentation, including Logfire.
Use Cases
- A Python developer can equip a Pydantic AI agent to investigate a failing test, edit code, and verify a fix.
- A research team can assemble an agent that searches the web, reads sources, and delegates focused subquestions.
- An application team can add persistent memory and conversation search when useful information must carry across sessions.
- A team running long agent workflows can combine context controls and persistence to manage runs that may span restarts or extended sessions.
What you need
Detected in the repository
- Python >=3.10 (from pyproject.toml)
- A test suite and automated checks on GitHub Actions
License in plain words
MITpermissive
- Commercial use: yes
- Modify and redistribute: yes
- You must keep: the copyright and license notice
- Share your changes: no
A summary, not legal advice: the LICENSE file is what applies.
Getting Started
For the package, install with uv add pydantic-ai-harness or pip install pydantic-ai-harness. For current development, documentation, and issue tracking, use the Pydantic AI repository. See the README for capability-specific extras and examples.
Alternatives
- lagent: Lagent is a general Python agent framework with tools and memory, while pydantic-ai-harness adds workspace, delegation, and other capabilities to Pydantic AI agents.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| pydantic-ai-harness | Python | MIT | 958 | Archived |
| lagent | Python | Apache-2.0 | 2.3k | Quiet |
Considerations
The repository is marked archived and read-only, even though its measured history shows recent commits, releases, and merged pull requests. That activity should not be taken as a sign that this repository remains the development or support destination: the README directs users to Pydantic AI for current code, documentation, issues, and pull requests. The package uses 0.x versioning, and its README warns that APIs may change between minor releases. Python 3.10 or newer is required; optional capabilities may have additional dependencies. The project is MIT-licensed.
Found this useful?
Share it with someone who would like pydantic-ai-harness.