DeepSeek Harness: Modular Open-Source Framework for AI Agents

LinkDeepSeekAnnouncement

DeepSeek has launched a developer preview of DeepSeek Harness, an open-source runtime framework for AI agents. Built on the Cordis kernel, it implements an architecture where all capabilities are plugins alongside an append-only event stream for full traceability.

Key points

  1. Modular plugin architecture: Models, tools, skills, sandboxes, storage, execution loops, scheduling, and UI are all plugins managed without altering core code.
  2. Cordis kernel: Handles mounting, unmounting, and dependency management between plugins via services and events.
  3. Comprehensive session logging: An append-only log records prompts, reasoning, tool outputs, and context injections, enabling trajectory inspection, replays, forking, and search.
  4. Four runtime presets: Standard mode for full coding workflows, Code mode using a TypeScript SDK to orchestrate tools, Minimal mode for benchmarking, and Creator mode for runtime inspection and authoring presets.
  5. Open source: Released under the MIT licence, installable via npm (npx @deepseek-ai/dsh web) or direct source from GitHub.

Why it matters

Building production-ready AI agents often suffers from rigid architectures and opaque decision paths. DeepSeek Harness provides a modular, configurable foundation that makes every step inspectable and reproducible through structured event streams.

Analysis

By framing the agent environment as a pluggable harness distinct from the underlying model, DeepSeek provides developers with granular control over agent capabilities. The separation into distinct modes—ranging from minimal benchmarking setups to multi-step TypeScript execution—addresses practical development needs. Given its developer preview status, the plugin ecosystem and API surfaces remain subject to ongoing iteration.

Mentioned links

Found in

Summary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.

Related entries

Similar in content, found by the AI search.

  • Repository:deepseek-ai/deepseek-harness

    DeepSeek Harness: Plugin-Based Open-Source AI Agent Framework

    DeepSeek Harness (dsh) is an open-source AI agent harness by DeepSeek AI. Built with TypeScript and Cordis, it uses an 'everything-is-a-plugin' architecture to provide modular agent capabilities and a Web UI.

    212.8kstarsTypeScript

    AI & LLMs· Tool

  • Article:Olimpiu Pop

    DeepSeek Harness: Modular Open-Source Runtime for AI Agents

    DeepSeek has released a developer preview of DeepSeek Harness (dsh) under the MIT licence. The execution runtime for autonomous artificial intelligence agents is built on the Cordis meta-framework using a micro-kernel architecture with interchangeable plugins.

    AI & LLMs· News

  • Repository:tonyd2wild/DeepSeek-Harness-Tools

    DeepSeek Harness Tools: Community Extensions for dsh

    DeepSeek-Harness-Tools serves as an unofficial index of add-ons for the open-source DeepSeek Harness (dsh). The collection brings together modular tools to equip local AI agents with web access, computer vision, and in-app browser control.

    14stars

    AI & LLMs· Tool

  • Repository:walkinglabs/learn-harness-engineering

    Learn Harness Engineering: Open-Source Course for Reliable AI Coding Agents

    The open-source project learn-harness-engineering by walkinglabs provides a project-based beginner curriculum on harness engineering for autonomous coding agents. It teaches how to build runtime environments, state management, verification routines, and control systems needed to operate AI agents reliably.

    14.8kstarsTypeScript

    AI & LLMs· Tutorial

Back to the knowledge base

Let's talk about your project

Get in touch

Locations

  • Mattersburg
    Johann Nepomuk Bergerstraße 7/2/14
    7210 Mattersburg, Austria
  • Vienna
    Ungargasse 64-66/3/404
    1030 Wien, Austria

Parts of this content were created with the assistance of AI.