Products / Release 0.1.0

Verxl on your
desktop.

Chat, Playground, Registry and API as one installed workspace. The same code, the same account, the same clearing path — on macOS, Windows, Linux and Android.

Choose your platform Open the web app API keys ↗
Choose a build4 platforms1 codebaseLive workspace
01 / Native

Download the workspace.

v0.1.0Release notes ↗Checksums ↗
⌘
macOS Your build
Apple Silicon · macOS 12+
DMG98 MB
Unsigned: right-click → Open
Download ↓
⊞
Windows Your build
Windows 10/11 · x64
EXE82 MB
Unsigned: SmartScreen may warn
Download ↓
$_
Linux Your build
x86_64 · AppImage
AppImage108 MB
chmod +x, then run
Download ↓
A
Android Your build
Android 7+ · debug APK
APK3.7 MB
Sideload; Play Store build later
Download ↓
iOS and any other platform

Install the same workspace from the browser. Safari: Share → Add to Home Screen.

Open web app ↗

Desktop builds are unsigned. macOS: right-click → Open. Windows: SmartScreen may warn. Android is a debug APK until Play Store signing. All shells load the live workspace, so updates ship with the web.

02 / API

Build against the clearing layer.

One OpenAI-compatible endpoint. 50 bps on compute. 99.5% to the provider. USDC settlement on Arbitrum. Start with a key, test in the playground, ship the same route.

Create API key ↗Open playground ↗
50 bpsclearing fee
99.5%to provider
USDCArbitrum
POST /v1/chat/completions
curl https://verxl.com/v1/chat/completions \
  -H "Authorization: Bearer $VERXL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4.1-flash",
    "messages": [{ "role": "user", "content": "Hello" }],
    "stream": true
  }'
03 / Harness

Bring your own harness.

Keep the editor, CLI or coding agent you already use. Point it at Verxl through the OpenAI-compatible base URL and select a routed model. Your harness stays yours; the clearing path handles access, metering and settlement.

No SDK requiredUse the standard OpenAI Chat Completions API.
StreamingToken streaming with usage and clearing receipt.
Model choiceQwen 3.5 4B, DeepSeek V4.1 Flash, MiMo 2.6 Flash.
Tool callingWhere the routed model supports it.
Create a key ↗Test a request ↗
YOUR SHELL
$ export OPENAI_BASE_URL="https://verxl.com/v1"
$ export OPENAI_API_KEY="vrxl_..."
$ your-harness --model deepseek-v4.1-flash
Base URLhttps://verxl.com/v1
ChatPOST /v1/chat/completions
StreamingSSE · data: [DONE]
AuthAuthorization: Bearer
BRING YOUR OWN TOOLSSource ↗
◇
Unsigned, checksummedSHA-256 checksums ↗
◇
Built by ActionsNative build pipeline ↗
◇
Same workspaceOpen the four-page app ↗