https://github.com/xyz2b/weave-ai-runtime
Composable AI runtime framework for building agent systems with tools, agents, skills, hosts, memory, workflow packages, and scenario packs.
https://github.com/xyz2b/weave-ai-runtime
agent-framework agentic-ai ai-assistant ai-coding ai-runtime autonomous-agents chatbot-framework code-generation coding-assistant conversational-ai dev-tools headless-ai hook-system llm-orchestration llm-tool-calling local-ai local-llm multi-agent-systems workflow
Last synced: 29 days ago
JSON representation
Composable AI runtime framework for building agent systems with tools, agents, skills, hosts, memory, workflow packages, and scenario packs.
- Host: GitHub
- URL: https://github.com/xyz2b/weave-ai-runtime
- Owner: xyz2b
- License: apache-2.0
- Created: 2026-05-08T16:04:45.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2026-05-27T15:25:24.000Z (about 2 months ago)
- Last Synced: 2026-05-27T17:14:20.942Z (about 2 months ago)
- Topics: agent-framework, agentic-ai, ai-assistant, ai-coding, ai-runtime, autonomous-agents, chatbot-framework, code-generation, coding-assistant, conversational-ai, dev-tools, headless-ai, hook-system, llm-orchestration, llm-tool-calling, local-ai, local-llm, multi-agent-systems, workflow
- Language: Python
- Homepage:
- Size: 3.06 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md
- Security: SECURITY.md
Awesome Lists containing this project
README
English | [简体中文](README.zh-CN.md)
# WeaveRT
Composable AI runtime framework for building agent systems with tools, agents, skills, hosts, memory, workflow packages, and scenario packs.
## What is WeaveRT?
WeaveRT is a runtime framework for building and operating agent systems.
It gives you a stable runtime core plus clear extension surfaces for tools, agents, skills, hosts, permissions, memory, workflow packages, and scenario packs.
It is not a single preset assistant app.
You can start with a small project-local workflow, then grow into coding, chat, or local-assistant products without rewriting the runtime model.
## Why WeaveRT?
- Build on a runtime, not a single prompt.
- Compose tools, agents, skills, and packages without hiding ownership boundaries.
- Keep host integration, permissions, and durable state explicit.
- Start with a minimal scaffold, then grow into richer workflows and apps.
## Quickstart
The default first run is the starter, not `examples/`.
Starter is the adoption path. Examples are the validation path.
From a published package index, the shortest path is:
```bash
python3 -m venv .venv
source .venv/bin/activate
python -m pip install weavert-starter weavert-testing
weavert-starter generate minimal-project ./my-weavert-app
cd my-weavert-app
python -m pip install -e .
python app.py
```
If you are editing WeaveRT itself from a local source checkout, use [docs/getting-started/install-from-source.md](docs/getting-started/install-from-source.md) and then follow the same starter flow.
Expected first-run anchors:
- `preset: ordinary-workflow`
- `assistant: The scaffold is ready...`
- `status: ok`
## Start Here
- Docs home: [docs/README.md](docs/README.md)
- First run: [docs/getting-started/quickstart.md](docs/getting-started/quickstart.md)
- Official starter path: [docs/getting-started/starter-scaffolds.md](docs/getting-started/starter-scaffolds.md)
- Runnable validation path: [examples/README.md](examples/README.md)
## Choose Your Path
- I want the smallest successful project -> [docs/getting-started/starter-scaffolds.md](docs/getting-started/starter-scaffolds.md)
- I want the first real project path -> [docs/guides/build-your-first-project.md](docs/guides/build-your-first-project.md)
- I want to understand the runtime model: read [docs/introduction/what-is-weavert.md](docs/introduction/what-is-weavert.md) and [docs/concepts/runtime-model.md](docs/concepts/runtime-model.md)
- I want to extend tools, agents, or skills: go to [docs/guides/add-a-tool.md](docs/guides/add-a-tool.md), [docs/guides/add-an-agent.md](docs/guides/add-an-agent.md), and [docs/guides/add-a-skill.md](docs/guides/add-a-skill.md)
- I want host, hook, or control-plane integration: go to [docs/guides/bind-a-host.md](docs/guides/bind-a-host.md), [docs/guides/extend-the-control-plane.md](docs/guides/extend-the-control-plane.md), and [docs/guides/register-hooks.md](docs/guides/register-hooks.md)
- I want to validate real workflows: use [examples/README.md](examples/README.md)
- I maintain the repository: use [docs/maintainers/repository-layout.md](docs/maintainers/repository-layout.md) and [docs/maintainers/migration-notes.md](docs/maintainers/migration-notes.md)
## Examples
- Seam basics and validation path: [examples/README.md](examples/README.md)
- Ordinary coding workflow validation: [examples/README.md](examples/README.md)
- Advanced host-bound integration sample: [examples/apps/code_assistant/README.md](examples/apps/code_assistant/README.md)
Generate a starter first unless you are specifically evaluating framework seams or validation evidence.
## Architecture and Reference
- Architecture overview: [docs/architecture/overview.md](docs/architecture/overview.md)
- Request lifecycle: [docs/architecture/request-lifecycle.md](docs/architecture/request-lifecycle.md)
- RuntimeConfig reference: [docs/reference/runtime-config.md](docs/reference/runtime-config.md)
- Full docs index: [docs/README.md](docs/README.md)
## Status
WeaveRT is under active development.
The documentation is organized as a layered journey:
landing page -> getting started -> concepts -> guides -> architecture/reference/maintainers.
## Contributing
See [CONTRIBUTING.md](CONTRIBUTING.md) for setup, workflow, and documentation conventions.
Community expectations live in [CODE_OF_CONDUCT.md](CODE_OF_CONDUCT.md), and vulnerability reporting guidance lives in [SECURITY.md](SECURITY.md).
## License
Apache-2.0. See [LICENSE](LICENSE).