Why KsADK¶
KsADK is not another agent framework. It is a unified runtime layer for agents built with existing frameworks.
Most frameworks answer:
KsADK focuses on:
One-line Positioning¶
Build agents once. Run them anywhere.
KsADK is the Agent Runtime Platform for AI agents.
Keep building business logic with Google ADK, LangGraph, LangChain, or DeepAgents. Use KsADK for one CLI, local Web UI, OpenAI-Compatible API, Skill Runtime, Workspace, Sandbox, deployment, and OpenTelemetry observability path.
Why Not Just ADK¶
Google ADK focuses on agent modeling, tools, multi-agent collaboration, Session/Memory, local runs, and Web debugging.
KsADK does not replace ADK. KsADK lets ADK agents share the same local runtime, browser debugging UI, OpenAI-Compatible API, and deployment entrypoint as LangGraph, LangChain, and DeepAgents projects.
KsADK is useful when:
- your team has both ADK and LangGraph projects.
- you want one Web UI to debug agents built with different frameworks.
- you need to expose local agents through OpenAI-Compatible APIs.
- you need Kingsoft Cloud AgentEngine, Skill, Workspace, Sandbox, or deployment integration.
Why Not Just LangGraph¶
LangGraph is strong at graph-state orchestration, checkpointing, streaming, human-in-the-loop workflows, and the LangChain ecosystem.
KsADK does not rewrite LangGraph execution. It adds a runtime platform layer around LangGraph projects:
agentengine runfor local terminal interaction.agentengine webfor browser debugging./v1/responsesand/v1/chat/completionslocal protocols.- Skill Runtime, Workspace, and Sandbox toolsets.
- AgentEngine, Hermes, OpenClaw, and Serverless deployment entrypoints.
Why Not Just OpenAI Agents SDK¶
OpenAI Agents SDK is native to the OpenAI Responses API and provides agent orchestration, tool calling, handoffs, guardrails, and tracing.
KsADK is multi-framework and multi-runtime. It does not require your business agent to be written with the OpenAI Agents SDK. It wraps ADK, LangGraph, LangChain, and DeepAgents projects into one local runtime, debugging, OpenAI-Compatible API, and deployment workflow.
What KsADK Solves¶
| Problem | KsADK capability |
|---|---|
| Start a new agent project quickly | agentengine init, agentengine config, agentengine run |
| Debug streaming, attachments, and tool calls in a browser | agentengine web and the ksadk-web debugging UI |
| Use one invocation protocol across frameworks | /v1/responses, /v1/chat/completions |
| Add tools, workspace access, and isolated execution | Skill Runtime, Workspace tools, Sandbox tools |
| Deploy to remote runtimes | agentengine build, agentengine launch, Hermes/OpenClaw |
| Observe consistently | OpenTelemetry / OTLP tracing |
When You May Not Need KsADK¶
If you only have a single-file script that directly calls a model API and you do not need browser debugging, isolated tools, unified protocols, deployment, or observability, the framework itself may be enough.
KsADK becomes useful when you need one way to manage multiple agent projects.