# OpenWarp landing page (openwarp.zerx.dev)

# OpenWarp — Unlock custom AI providers for Warp

Community · Work in progress · Early development — no public release yet.

## Tagline

> Bring any AI model into your terminal. OpenWarp adds BYOP (Bring Your Own Provider) on top of Warp — 6 native API protocols via the genai adapter, your own models and prompts, first-class i18n.

```
git clone -b openWarp https://github.com/zerx-lab/warp
```

Compatible with major providers: OpenAI, Anthropic, DeepSeek, Qwen, Ollama, Groq.

Stat block on the page: ∞ pluggable providers, 2 built-in locales, AGPL open-source license, 100% local credentials.

## How it works

> Three steps. Your terminal. Your AI. Keep the full Warp experience — only the AI layer is opened up. Your keys, your models, your prompts.

01 — **Plug in any provider.** Pick the API protocol, paste a Base URL and API key in settings — switch freely across OpenAI / Anthropic / Gemini / Ollama / DeepSeek (6 native protocols), credentials stay on-device.

02 — **Author dynamic prompts.** A minijinja template engine renders your system prompt from the cwd, locale, and role — context-aware on every call.

03 — **Use it in the terminal.** Switch models, chat, and complete commands the same way you use Warp — except now you own the stack end-to-end.

## Sample custom-provider config

```toml
# ~/.config/openwarp.toml
name = "DeepSeek (Personal)"
base_url = "https://api.deepseek.com/v1"
api_key  = "sk-……"
model    = "deepseek-r1"
```

## Core features

- **BYOP custom providers** — 6 native API protocols via the genai adapter. Mix any Base URL / API Key / Model.
- **6 native API protocols, not just an OpenAI-compat shim**
- **Multi-turn reasoning passthrough**: DeepSeek `reasoning_content` / Claude `thinking` / Gemini.
- **minijinja-powered system prompt templates**.
- **Privacy first** — Cloud Agent / Computer Use disabled by default. Credentials stay local.
- **Local credentials**: 100%, no cloud upload, no telemetry, zero credential leak.
- **minijinja prompt templates** — render instructions dynamically from context.
- **Multilingual UI** — first-class Chinese & English. Community can grow the locale set (Japanese, Spanish previewed).
- **Warp-native experience** — continuously merging upstream — keep blocks, AI commands, workflows, keymaps, themes.
- **Open source** — AGPL / MIT dual license, mirroring upstream Warp.

## Custom Providers

> Configure once, every model unlocked. OpenWarp speaks 6 native API protocols via the genai adapter: OpenAI / OpenAI Responses / Anthropic / Gemini / Ollama / DeepSeek — protocol is explicit, no model-name guessing, keys go straight to the provider.

- ✓ 6 native API protocols, not just an OpenAI-compat shim
- ✓ Multi-turn reasoning passthrough: DeepSeek reasoning_content / Claude thinking / Gemini
- ✓ minijinja-powered system prompt templates
- ✓ Credentials stored locally, requests go straight to the provider endpoint

Sample provider entry:

```
Provider name:    DeepSeek (OpenAI-compat)
API protocol:     OpenAI
Base URL:         https://api.deepseek.com
Request endpoint: POST /v1/chat/completions
API key:          sk-…
Default model:    deepseek-reasoner
```

System prompt template example:

```jinja
{% if user.role %}
You are an expert {{ user.role }}.
{% endif %}
Reply in {{ locale }}.
```

Provider catalog tiles shown: OpenAI, Anthropic, DeepSeek, Qwen, Ollama, Groq, Together, Mistral, Gemini, LM Studio, OpenRouter, Azure.

## FAQ

**How is OpenWarp related to Warp the company?**
> OpenWarp is a community fork of Warp's open-source code. It is not affiliated with Warp Inc. and follows the upstream AGPL / MIT dual license.

**Will my API keys be uploaded?**
> No. Custom provider credentials live only in your local config file. OpenWarp talks directly to the Base URL you configure, with no relay.

**Which providers are supported?**
> OpenWarp ships a genai-based multi-protocol client: OpenAI / OpenAI Responses / Anthropic / Gemini / Ollama / DeepSeek run as native protocols. OpenAI-compatible endpoints (Qwen / Groq / Together / OpenRouter / SiliconFlow / LM Studio, etc.) plug in by selecting the OpenAI protocol and pointing Base URL at them.

**Will I keep getting upstream Warp updates?**
> Yes — we continuously merge from upstream Warp while layering BYOP and i18n on top.

## Build / install

```
~/dev$ git clone -b openWarp https://github.com/zerx-lab/warp
1  cargo build --release
2  ./target/release/openwarp
3  Add a custom provider in Settings
```

## Footer

A community fork on top of Warp. Not affiliated with Warp. © 2026 OpenWarp · AGPL-3.0 / MIT.

Site built with Astro · Tailwind · Framer Motion. Domains: openwarp.zerx.dev (current), openwarp.dev (mentioned).

## Repo metadata (from GitHub API at fetch time)

- `zerx-lab/warp`, fork of `warpdotdev/warp`
- Created 2026-04-29; pushed 2026-05-02
- 422 stars, 23 forks, 5 open issues, no releases yet
- License: AGPL-3.0 (per repo); site mentions AGPL/MIT dual
- Primary language: Rust
- Branch: `openWarp` (capital W)