DeepSeek Harness
Open-source agent runtime — everything is a plugin
AICoding score
Quick Verdict
7.8 / 10 — The most flexible open-source agent runtime in 2026 — MIT licensed, "everything is a plugin", with a 700+ plugin ecosystem that grew in days.
DeepSeek Harness is a kit, not a product: you assemble the agent you want by stacking plugins — models, tools, vision, sandboxes and even the UI are swappable. The payoff is zero lock-in and a genuinely open loop; the cost is v0.1-preview instability and assembly work. If you want a turnkey coding agent, use Claude Code or Codex. If you want to own the runtime and build your own agent platform, dsh is the strongest base in 2026.
Use dsh if…
- You want an open, MIT-licensed agent loop with no vendor lock-in
- You want to add vision / image recognition via plugins (ModLens, open-design)
- You run multi-agent teams and want traceable, replayable runs
- You want to migrate from Claude Code without losing skills
Skip dsh if…
- You need production stability this month — v0.1 preview, breaking changes
- You want a turnkey product with first-party support (Claude Code / Codex)
- You dislike assembling and maintaining your own stack
Core Features
Everything Is a Plugin
The core design principle and the reason the ecosystem exploded. The model adapter, tool registry, session log, sandbox, approval policy and even the UI are all plugins over the Cordis kernel — no privileged core, extensions mount as siblings and unwind on unload. You replace any capability from a config file without forking source.
53 Built-in Tools
bash, pwsh, edit, read, write, glob, grep (bundled ripgrep), lsp (language server protocol), subagent + subagent_fork, web_fetch, persistent terminal sessions, and the cordis_* family. The catalog is generated by actually booting each plugin and reading its schemas, with a build guard that fails if a tool package is missing.
Vision & Image Recognition via Plugins
DeepSeek's V4 models are text-only, but ModLens and dsh-vision-toolkit convert images into structured evidence (text, layout, entities) before reasoning, and open-design (87k+ stars) turns the harness into a local design workbench that generates and previews prototypes, webpages, slides, images and video. Vision is one plugin away, not a model limitation.
Traceable by Design
An append-only session log records everything the model saw — system prompts, reasoning, tool calls, subagent scheduling, raw streaming chunks — replayable down to the failed step. The invariant is "model-visible means logged". For debugging agent runs and compliance auditing, this is the strongest replay story of the harness category.
Multi-Agent Teams & Orchestration
dsh-agent-teams promotes an agent to team leader, splits tasks, sets dependencies, and lets members message each other, with the Web UI showing real-time status of who is working. Combined with the 4 runtime modes (Standard, Code/PTC, Minimal, Creator), it is a full orchestration surface.
Migration & Long-Term Memory
dsh-plugin-claude-bridge imports Claude memories, skills and CLAUDE.md; dsh-claude-move migrates whole sessions so you continue the conversation in Harness. dsh-memory-evolve adds cross-session memory that persists project conventions and decisions across sessions — turning the agent from stateless chat into a project-aware assistant.
Third-Party Plugin Picks
Vision & Design
Local design workbench — generate and preview prototypes, webpages, slides, images and video; DSH-native runtime adapter. The must-have for design-adjacent work.
dsh install open-design Adds image paste, OCR and layout understanding to text-only models — converts screenshots into structured evidence before reasoning.
dsh install modlens Vision tool suite for image input and analysis — pairs with ModLens for a full vision stack.
dsh install dsh-vision-toolkit Teams & Memory
Promote an agent to team leader, split tasks, set dependencies, let subagents message each other — live status in the Web UI.
dsh install dsh-agent-teams Cross-session long-term memory: persists project conventions, architecture decisions and pitfalls; evolves skills in the background.
dsh install dsh-memory-evolve Context database — unified memory, knowledge, RAG and Skills for the agent across sessions and tools.
dsh install openviking Migration & Integration
Imports Claude Code memories, skills and configuration so you keep your Claude setup inside dsh.
dsh install dsh-plugin-claude-bridge Full migration of Claude sessions — session, memory, skills and CLAUDE.md — to continue the same conversation in Harness.
dsh install dsh-claude-move Manage GitHub directly from the conversation — issues, PRs, merges without leaving the agent.
dsh install dsh-github-connector UI & Workbench
Desktop client — no Node.js required, download and launch the harness as a native app.
download from repo Web UI plugin family: task boards, Git graph, right-side panel, mobile support and live stats.
dsh install dsh-web-ui Mini-IDE sidebar — file management, code editing, terminal, Git panel and background tasks in one place.
dsh install dsh-better-sidebar Workflow & Tools
Generates verifiable architecture and flow diagrams from your codebase — great for docs and planning.
dsh install archify Codex-style @file reference — type @login.ts in the prompt and inject the file content directly.
dsh install dsh-at-file Rolls back a broken agent context — the undo button agents were missing.
dsh install dsh-undo Record a sequence of actions once, replay it later — great for demos and reproducible setup.
dsh install dsh-record-replay Pricing
The entire runtime — plugin kernel, 53 tools, all runtime modes, Web UI/TUI. No paid tier exists behind the free one; the project is MIT and self-hostable on your own machines.
- MIT open source
- All runtime modes
- Web UI + TUI
- No paid tier
Point dsh at any OpenAI-compatible provider — DeepSeek, Claude, GPT, Gemini — or a local Ollama model. No subscription, no vendor lock-in, data stays with you.
- Any OpenAI-compatible API
- Local models via Ollama
- No subscription
- Zero vendor lock-in
The recommended default for routine agent work: $0.22 input / $0.66 output per M off-peak, with weekends now billed all-day off-peak and heavy cache discounts. A daily agent workload lands in low single-digit dollars per month.
- Value open API
- 99% cache-hit friendly
- Weekends off-peak all day
- ~$1-5/mo typical
TerminalBench 2.1 at 87.9 — the first open model to top the terminal benchmark. $1.32/$3.96 at peak, half-price off-peak. Use it for hard refactors and planning; route routine steps to V4-Flash via plugin.
- 87.9 TerminalBench 2.1
- 1M context
- Off-peak half price
- OpenAI Responses format
* Estimates from published usage reports & community benchmarks. Your API cost depends on model choice and volume. Local models via Ollama cost $0.
DeepSeek Harness vs The Field
dsh vs Claude Code
Open loop vs closed verticalClaude Code is a closed vertical: Anthropic's loop around Anthropic's models. dsh is the opposite — MIT-licensed, plugin-driven, model-agnostic, and it can even migrate your Claude sessions (dsh-plugin-claude-bridge). Claude Code wins on maturity and first-party support; dsh wins on freedom, cost control and no lock-in. Most teams keep both: Claude Code for production, dsh to explore.
dsh for openness · Claude Code for maturitydsh vs Codex CLI
Both open source, opposite architectureCodex CLI is Apache-2.0 but architecturally a closed vertical around OpenAI — the loop is fixed, extensibility is via plugins on OpenAI's terms. dsh makes the entire loop replaceable and is model-agnostic. Codex is more stable and better integrated with ChatGPT plans; dsh is the assembly kit. For developers who want to own the runtime: dsh.
Codex for stability · dsh for ownershipdsh vs OpenHands
Open agent framework, different focusOpenHands is a turnkey open-source agent you run, with a UI and workflow out of the box. dsh is a plugin meta-framework you assemble — more setup, far more flexibility, and a much faster-growing plugin ecosystem (700+ vs OpenHands' extension model). OpenHands if you want working software today; dsh if you want to build your own agent platform.
OpenHands for turnkey · dsh for customizationWho dsh Is Best For
Escape vendor lock-in
MIT license + plugin architecture + model-agnostic adapters mean no single vendor owns your agent loop. Swap models, sandboxes and UI without changing the harness — the anti-closed-vertical choice.
Building an agent product
Use dsh as a composable base instead of forking Claude Code or wrapping Codex. Cordis dependency injection and reversible effects make plugins the extension model for your product.
Vision & design work
Plug in ModLens or open-design to let the agent read UI screenshots, generate prototypes and preview webpages/slides — closing the text-only gap of DeepSeek's V4 models.
Compliance & audit
The append-only, token-level-replay session log is the strongest audit story in the category — model-visible means logged, so every agent action is reconstructible.
Multi-agent orchestration
dsh-agent-teams turns the harness into a team manager: leader splits tasks, subagents message each other, and the Web UI shows live status. Built for the 2026 multi-agent standard.
Claude Code migration
If you are leaving Claude Code, dsh-plugin-claude-bridge and dsh-claude-move import skills, memory and sessions so you pick up where you left off — no rewrite.
Getting Started
Install & launch
Run npx @deepseek-ai/dsh web — the Web UI starts on http://127.0.0.1:3080. No install script, no accounts, works from any directory.
Connect a model
Add a DeepSeek API key (or any OpenAI-compatible key) in the settings. For local, point it at Ollama. Model adapters are plugins — swapping providers is a config change.
Discover plugins
Use dsh-find-plugin or browse dsh-index.xlings.org / awesome-dsh-plugins to find bundles. Install with a single command per plugin.
Add vision (optional)
Install modlens or open-design so the agent can read UI screenshots and generate/preview designs — the most common "make it actually useful" upgrade.
Pick your mode
Standard for daily work, Code (PTC) to keep intermediate tool data out of context and save tokens, Minimal for benchmarks, Creator to let the agent inspect and extend its own runtime.
Migrate or go multi-agent
Import Claude Code skills with dsh-plugin-claude-bridge, or install dsh-agent-teams to orchestrate a team of subagents from one conversation.
FAQ
Is DeepSeek Harness free?
Yes — fully free and MIT licensed. The harness itself costs $0 forever, on every machine. You only pay for model API usage (or run local models via Ollama for $0). "Everything is a plugin" means even the UI is replaceable, so there is no paid tier hiding behind the free one — the project is MIT open source with no commercial licensing gate.
Is DeepSeek Harness a model?
No. dsh is an agent runtime (a harness) — the layer that runs models. It is DeepSeek's answer to Claude Code and Codex CLI, but instead of bundling a model, it lets you plug in any model: DeepSeek, Claude, GPT, Gemini, or a local Ollama model. "Everything is a plugin" applies to models too — swap the adapter from a config file without changing the harness.
How is DeepSeek Harness different from Claude Code / Codex?
Three structural differences. (1) It is open source (MIT) and the agent loop itself is open — Claude Code and Codex are closed verticals around their own models. (2) Everything is a plugin: model adapter, tools, session log, sandbox, even the UI are swappable plugins over the Cordis kernel. (3) It is a kit, not a turnkey product — more setup work, far more freedom. Same category, opposite design philosophy.
How many DeepSeek Harness plugins are there?
Over 700 public repos carry the dsh-plugin topic (within 24 hours of release), and the index site dsh-index.xlings.org aggregates them into a searchable catalog. They range from productivity (multi-agent teams, cross-session memory, Claude Code migration, vision) to themes and games. The community compares the ecosystem to Minecraft: dsh is the survival mode, plugins are mods, awesome-dsh-plugins is CurseForge.
Can DeepSeek Harness see images / do vision?
Not natively for the DeepSeek V4 family (text-only), but the plugin model fixes it: ModLens and dsh-vision-toolkit extract text, layout and entities from images into structured evidence before reasoning, and open-design (a local design workbench) lets the agent generate and preview prototypes, webpages, slides, images and video. Vision is one plugin away rather than a model limitation.
Can I migrate from Claude Code to DeepSeek Harness?
Yes — two plugins handle it. dsh-plugin-claude-bridge imports Claude memories, skills and configuration files; dsh-claude-move performs a full migration of sessions, memory, skills and CLAUDE.md so you continue the same conversation inside Harness. This mirrors the wider 2026 trend of skills becoming portable across agent ecosystems.
Is DeepSeek Harness stable for production?
Not this month. It is a v0.1 developer preview that explicitly warns of compatibility-breaking changes. The plugin ecosystem is moving extremely fast (700+ repos in days), so pin versions and treat it as an assembly base. For production-critical work, pair it with budget models for routine tasks and keep a stable fallback tool until the preview matures.
What models can DeepSeek Harness use?
Any OpenAI-compatible API: DeepSeek (V4-Pro, V4-Flash), Claude, GPT, Gemini, or local models via Ollama. The recommended default is DeepSeek V4-Flash for budget and V4-Pro for complex work — the same routing pattern teams now use across all harnesses: cheap model for routine steps, frontier model for hard ones.
How much does it cost to run?
The tool is free; the bill is model usage. DeepSeek V4-Flash is $0.22 / $0.66 per M tokens off-peak (and weekends are now all-day off-peak), with V4-Pro at $1.32 / $3.96 and cached input at $0.44 — far below subscription pricing. Local models via Ollama cost $0. A typical daily agent workload with V4-Flash lands in low single-digit dollars per month.
Does DeepSeek Harness have a GUI?
Yes, several. The official Web UI (dsh-web-app) runs on http://127.0.0.1:3080, dsh-TUI gives a full-screen terminal with status meters, and community clients like deepseek-harness-desktop (10k+ stars) package the whole thing without requiring Node. Because UI is a plugin, you are not stuck with one interface.