Ebony-Vinyl/dsh-our-free-model
Adds free model lanes to DeepSeek Harness — DeepSeek V4.1 Flash, Kimi K3 and others — with no login, sign-up or API key, plus a local OpenAI-compatible gateway, fallback routing and a usage dashboard.
Install
via GitHub · GitHub source
Installing from GitHub runs the project's build script, which pnpm blocks until you allowlist it — run the command once and pnpm prints the exact key to add under `allowBuilds` in ~/.dsh/profiles/web/pnpm-workspace.yaml.
The desktop app keeps its own profile, desktop, and only the dsh command it ships may change it — the npm-installed dsh refuses that profile. Install the command from the app menu (Manage dsh Command…), open Desktop once, quit it fully, run the line below, then reopen Desktop.
via GitHub · GitHub source
Installing from GitHub runs the project's build script, which pnpm blocks until you allowlist it — run the command once and pnpm prints the exact key to add under `allowBuilds` in ~/.dsh/profiles/desktop/pnpm-workspace.yaml.
Our install check ran in a web profile on npm's dsh, not in the desktop app.
What happened when we ran it
Installed cleanly when we ran it
Every command here is run in a throwaway container against a clean profile, and the result is whatever the harness recorded — not a guess from the source. Last run 6d agoon DeepSeek Harness 0.2.1-rc.5.
Show it in your README
For maintainers: the badge serves this listing's latest sandbox verdict, so a re-run updates it on its own — and it links readers to the full result here.
Due diligence
Security
No build script was blocked in our run.
Source: our sandbox install on DeepSeek Harness 0.2.1-rc.5, 6d ago.
package.json declares no preinstall, install or postinstall script.
Source: package.json on GitHub, read at our last sync. prepare is not recorded; it runs only on a GitHub install, and shows up under build scripts when pnpm blocks it.
shell commands (bash, exec, spawn …)
Source: a keyword match on the README — a hint about what the plugin touches, not an analysis of its code.
Not an audit. These are the record of one install and two text scans; nobody has read this plugin's code on your behalf. Plugins run with your agent's permissions.
What it is — Integrates free model lanes into DeepSeek Harness for models such as DeepSeek V4.1 Flash and Kimi K3 with no login required.
Who it is for — If you need free access to DeepSeek V4.1 Flash models in DeepSeek Harness for your tasks, this plugin fits. If your work relies solely on existing model providers, installation is unnecessary.
Watch out — Sandbox testing confirmed success after installing into a new profile with harness registration. The plugin will execute shell commands. No obvious issues were found.
The verdict — I would install it for the local OpenAI-compatible gateway enabling other tools, but only when free models are essential and network access is stable.
Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.
dsh-our-free-model is a DeepSeek Harness plugin that adds free model lanes, including DeepSeek V4.1 Flash and Kimi K3, without requiring an account or API key. DeepSeek Harness loads the plugin as an llm-dependent module. It refreshes the model list from OpenCode’s Zen gateway, probes availability from the local network exit, and exposes models through the model selector. The plugin also provides a local OpenAI-compatible gateway for /v1/models, /v1/chat/completions and /v1/responses, with streaming and non-streaming requests.
Usage measurements remain on the local machine and feed the dashboard’s token, latency and speed views. The plugin can also receive repository-hosted announcements and perform application upgrades with SHA-256 verification, backup, atomic replacement and rollback. DeepSeek Harness can load the model lane without a web server; the dashboard registers later when the web service becomes available.
This is intended for developers who want models available without managing upstream credentials, or who need a local OpenAI-compatible endpoint for other tools. It is a poor fit where requests must not leave the machine: the stated upstream source is OpenCode’s Zen gateway. Desktop hosts have additional profile constraints, and managed installations disable the plugin’s in-app upgrades, announcement feed and hot reload. Its terminal surface is the documented dsh and node command set used for installation, diagnostics and profile checks.
DeepSeek Harness loads the plugin in the selected profile, with llm as a hard dependency. The settings page exposes these documented options:
region-limited modelsThe forwarding service is disabled by default. When enabled, it provides GET /v1/models, POST /v1/chat/completions and POST /v1/responses. The local dashboard records usage data on the machine and does not upload it.
The README documents dsh plugin --profile web add for adding the package to a profile, with the actual profile name substituted for web. It also documents dsh --profile <name> --dump-config for checking the bundle configuration. A desktop self-check can load /sidecar/dist/lib/profile-upgrade.js with node and call planProfileUpgrade and assertProfileStartup.
For a real-directory desktop installation, the package is placed at node_modules/dsh-our-free-model/; dsh.profile.bundles must contain dsh-our-free-model, and dependencies must contain the current package version. The README identifies that version as 1.4.4.
The model collection, context length and capabilities are refreshed from the upstream gateway rather than stored as a static snapshot. A model rejected as unavailable or returning 404 is removed from the selector; gateway errors, rate limits, timeouts and network failures do not make that determination. Region-blocked models are grouped separately. Light, Balanced and Deep apply output budgets of 2,048, 8,192 and the model maximum, with doubled lower budgets for models whose reasoning cannot be disabled.
The upstream source is OpenCode’s Zen gateway, so requests depend on the local network route and may be region-limited. Desktop profile startup rejects arbitrary symlinks and junctions. Managed bundles must set distribution: 'managed'; this disables in-app upgrades, announcements and hot reload. The EAC channel exists only in the specified desktop hosts, not in command-line hosts.
Written from the project's own documentation and kept in sync with it. Where the two disagree, the source is authoritative — read the README on GitHub
Same category
zouyuxuan122
在 dsh 里装上这个插件即可，无需登录、注册或填 API Key，就能使用包括 Muse Spark 1.3、MiMo V2.6 在内的前沿模型——完全免费，不限量。 All you do is install this plugin in dsh: no login, no sign-up, no API key — the frontier models are just there, Muse Spark 1.3 and MiMo V2.6 among them. Completely free, with no usage cap.
Installed cleanly when we ran it
HunYiSage
DeepSeek Harness 破甲：让所有模型都能破甲，不同模型可换不同提示词；默认提示词面向国模「小码酱」。Jailbreak for every model — swap prompts per model. 求 Star 收藏 ⭐ 如要二开，请查看署名要求部分
Installed cleanly when we ran it
Use ChatGPT (Codex), Claude, and Grok subscriptions as DeepSeek Harness LLM providers, with Settings login, model catalogs, usage, plus image_generate, video_generate, and x_search tools.
Installed cleanly when we ran it
Mars-Sea
Unofficial Command Code LLM provider: registers a `commandcode` route with a live model catalog and reasoning-effort support.
Installed cleanly when we ran it
cv-superding
Adds a deepseek-web provider that uses chat.deepseek.com web models in DSH, with browser login capture, PoW request signing, SSE streaming, and prompting-based tool calls.
Installed cleanly when we ran it
aosi526
Merge every locally signed-in WorkBuddy account into DeepSeek Harness as one auto-failing-over model pool (multi-account rotation, live credits, and a model catalog showing per-model credit multipliers and free/night-discount tags).
Installed cleanly when we ran it