Edge Compute
Contact Sales Log in Sign up
Developers

Build on
Edge Compute.

CLI, Python SDK and REST API — everything you need to provision Canadian GPU compute from code, scripts, CI pipelines or AI agents.

Install CLI API reference
⌨️

CLI

Search, launch, SSH into and stop GPUs from your terminal. One file, no dependencies beyond Python 3.

Learn more
🐍

Python SDK

The same file is an importable SDK: from edge import EdgeClient. Five lines to a running GPU.

Learn more
🔌

REST API

Plain HTTPS + JSON for every operation, with an OpenAPI spec and interactive docs.

Open API docs
Get started in seconds

From zero to a running GPU

CLI

terminal
$ curl -fsSL https://clearlayer.ca/cli/install.sh | sh
$ edge set-key YOUR_API_KEY
$ edge search --gpu "RTX 5090" --sort value
ID  GPU       VRAM   TFLOPS  CPU  RAM    DISK     CUDA  RELIAB  PRICE
12  RTX 5090  32 GB  104.8   15   53 GB  1750 GB  13.0  99.8%   $0.79/h
$ edge create 12 --model qwen3-8b
instance 42 queued (vLLM (OpenAI-compatible API), Qwen3 8B).
$ edge show 42      # SSH, Jupyter, API URL + key
$ edge ssh 42
$ edge stop 42

Python SDK

python
# curl -fsSL https://clearlayer.ca/cli/edge.py -o edge.py
from edge import EdgeClient

c = EdgeClient(api_key="YOUR_API_KEY")
gpu = c.search_offers(gpu="RTX 5090", sort="price")[0]
inst = c.create_instance(
    gpu["id"],
    ssh_pubkey=open("/home/me/.ssh/id_ed25519.pub").read(),
    model="qwen3-8b",
)
print(c.get_instance(inst["id"])["app_url"])
c.stop_instance(inst["id"])
REST API

Endpoints

Base URL https://clearlayer.ca/api/v1 · auth header Authorization: Bearer YOUR_API_KEY

MethodPathAuthWhat it does
GET/offers—Search GPUs (gpu, min_vram, max_price, min_cpu, min_ram, min_disk, min_cuda, min_driver, min_reliability, pcie_gen, sort)
GET/templates—Template catalogue
GET/models—Model library
GET/accountkeyCredit, spend rate and runway
GET/instanceskeyYour active instances (?all=true for history)
POST/instanceskeyLaunch: gpu_id, ssh_pubkey, plan, template or model, max_hours, hf_token
GET/instances/{id}keyStatus and connection details (SSH, Jupyter, app URL, app token)
DELETE/instances/{id}keyStop and delete
curl -H "Authorization: Bearer $EDGE_API_KEY" https://clearlayer.ca/api/v1/account

Interactive API reference

Start building

Create an API key and launch your first GPU in a minute.

Get API key Read the docs