volcengine/veadk-python
An open-source kit for agent development, integrated the powerful capabilities of Volcengine.
What it solves
VeADK is an open-source development kit designed to simplify the creation and deployment of AI agents. It provides a structured framework for integrating Volcengine capabilities, managing agent topologies, and connecting agents to communication channels like Feishu.
How it works
The kit allows developers to build agents using a high-level Agent class and a Runner for execution. It integrates with Volcengine's ARK API for model providers (such as Doubao) and supports a configuration-driven approach via config.yaml for managing API keys and model settings.
For deployment, it includes an AgentKit application factory that provides a bundled Web UI, health checks, and metadata endpoints. It also features extensions, such as the FeishuChannelExtension, which maps chat IDs and user IDs to session IDs to enable memory and tracing within Feishu conversations.
Who it’s for
Developers building AI agents who want a streamlined way to integrate Volcengine's ecosystem, deploy agents with a built-in UI, and collect user feedback via like/dislike controls for evaluation.
Highlights
- Integrated Ecosystem: Direct integration with Volcengine ARK and the AgentKit application factory.
- Deployment Tools: Bundled Web UI and health check endpoints for easier runtime management.
- Deployment Flow: Support for cloud image builds with diagnostic logs for Dockerfile failures.
- Channel Extensions: Built-in bridging for Feishu bots to maintain session memory and tracing.
- Evaluation Loop: Server-side feedback collection (like/dislike) to create good and bad case evaluation sets.
Related
- Project
- Project
- Project
- Project
- Project