Vast.ai

https://docs.vast.ai
Admin
Vast.ai is a marketplace for affordable GPU cloud computing that allows users to spin up GPU...
Tokens:763,773
Snippets:6,760
Trust Score:9.8
Update:1 week ago
Show doc for...
Tokens:
Raw
### vastai start instances example

Source: https://docs.vast.ai/cli/reference/start-instances

Example of starting instances by providing their IDs.

```bash
vastai start instances <IDS>
```

--------------------------------

### vastai start instance examples

Source: https://docs.vast.ai/cli/reference/start-instance

Two examples: start all instances returned by 'vastai show instances -q', and start a specific instance by ID (329838).

```bash
vastai start instances $(vastai show instances -q)

```

```bash
vastai start instance 329838

```

--------------------------------

### Example: start instances with VastAI client

Source: https://docs.vast.ai/sdk/python/reference/start-instances

Example usage of the VastAI client to start an instance with ID 12345 and print the result.

```python
from vastai import VastAI

client = VastAI(api_key="YOUR_API_KEY")
result = client.start_instances(ids=12345)
print(result)

```

--------------------------------

### Example output with multiple kernel metapackages

Source: https://docs.vast.ai/host/upgrade-kernel

Example output showing multiple kernel metapackages installed. All of them are upgraded together.

```text
linux-generic-hwe-22.04
linux-generic-hwe-22.04-edge

```

--------------------------------

### Setup Ralph environment and dependencies

Source: https://docs.vast.ai/examples/ai-agents/overnight-ralph-loop

Sets up a Python virtual environment, installs the Vast CLI, Aider, and pytest, configures the Vast API key, and clones the Ralph repository. Run these commands in a terminal before starting the agentic loop.

```bash
# Create a virtual environment
python3 -m venv ralph-env
source ralph-env/bin/activate

# Install Vast CLI and Aider
pip install --upgrade vastai aider-chat pytest
vastai set api-key <your-vast-api-key>

# Clone Ralph
git clone https://github.com/snarktank/ralph.git
cd ralph

```

--------------------------------

### Example apt-cache policy output

Source: https://docs.vast.ai/host/upgrade-kernel

Example output from apt-cache policy showing the installed and candidate kernel versions.

```text
linux-generic:
  Installed: 5.15.0.25.27
  Candidate: 5.15.0.190.169
  Version table:
     5.15.0.190.169 500
        500 http://archive.ubuntu.com/ubuntu jammy-updates/main amd64 Packages
        500 http://security.ubuntu.com/ubuntu jammy-security/main amd64 Packages
 *** 5.15.0.25.27 500
        500 http://archive.ubuntu.com/ubuntu jammy/main amd64 Packages
        100 /var/lib/dpkg/status

```

--------------------------------

### dstack server startup output

Source: https://docs.vast.ai/examples/serving-infrastructure/dstack-vllm

Example output from starting the dstack server, showing the admin token and server URL.

```text
╭━━┳━━┳━┳╮╭┳━━┳━╮
┃━━┫┃━┫╭┫╰╯┃┃━┫╭╯
┣━━┃┃━┫┃╰╮╭┫┃━┫┃
╰━━┻━━┻╯╱╰╯╰━━┻╯

INFO     Applying ~/.dstack/server/config.yml...
INFO     Configured the main project in ~/.dstack/config.yml
INFO     The admin token is YOUR_ADMIN_TOKEN
INFO     The dstack server 0.19.40 is running at http://127.0.0.1:3000

```

--------------------------------

### Install Vast.ai plugin in Claude Code

Source: https://docs.vast.ai/guides/get-started/agents

Add the marketplace and install the plugin. Then run `/vastai:setup` once to store your API key, register your SSH public key, and verify the credential.

```shell
/plugin marketplace add vast-ai/vast-claude-plugin
/plugin install vastai
```

--------------------------------

### Set up MCP backend locally

Source: https://docs.vast.ai/examples/mcp/dr-tulu

Clones the dr-tulu repository, installs dependencies with uv, sets API keys, and starts the MCP backend on port 8001.

```bash
# Clone the repository
git clone https://github.com/rlresearch/dr-tulu.git
cd dr-tulu/agent

# Install uv if needed
curl -LsSf https://astral.sh/uv/install.sh | sh

# Install dependencies (uv sync handles venv creation automatically)
uv sync

# Configure API keys
export SERPER_API_KEY=your_serper_key
export JINA_API_KEY=your_jina_key

# Start the MCP backend
uv run python -m dr_agent.mcp_backend.main --port 8001
```

--------------------------------

### Example startup logs from vLLM

Source: https://docs.vast.ai/examples/ai-agents/browsesafe

Example logs showing successful model loading. The 'Application startup complete.' line indicates the server is ready.

```text
INFO vLLM API server version 0.13.0
INFO Resolved architecture: Qwen3MoeForCausalLM
INFO Using max model len 16384
INFO Starting to load model perplexity-ai/browsesafe...
INFO Using FLASH_ATTN attention backend
INFO Application startup complete.

```

--------------------------------

### Example /root/onstart.sh for SSH instances

Source: https://docs.vast.ai/guides/reference/faq/instances

Place startup commands in /root/onstart.sh to run automatically on container startup for SSH instances. This example changes to /workspace and resumes training.

```bash
# Example /root/onstart.sh
#!/bin/bash
cd /workspace
python train.py --resume
```

--------------------------------

### Check instance status and SSH in

Source: https://docs.vast.ai/examples/ai-agents/autoresearch

Waits for the instance to be ready and then SSHes into it. The template installs everything on first boot (~10 minutes).

```bash
vastai show instances
ssh -p PORT root@HOST_IP

```

--------------------------------

### Install Cohere client

Source: https://docs.vast.ai/examples/embeddings/serving-rerankers-vllm

Installs the Cohere Python client, required for the /rerank endpoint example.

```bash
pip install --upgrade cohere

```

--------------------------------

### Example PROVISIONING_SCRIPT Bash Script

Source: https://docs.vast.ai/guides/templates/advanced-setup

Bash script that runs on first boot to set up the environment. Installs packages, downloads files, sets up services, and reconfigures the instance portal.

```bash
#!/bin/bash

#
cd /workspace/
# Cause the script to exit on failure.
set -eo pipefail

# Activate the main virtual environment
. /venv/main/bin/activate

# Install your packages
pip install your-packages

# Download some useful files
wget -P "${WORKSPACE}/" https://example.org/my-application.tar.gz
tar xvf ${WORKSPACE}/my-application.tar.gz"

# Set up any additional services
echo "my-supervisor-config" > /etc/supervisor/conf.d/my-application.conf
echo "my-supervisor-wrapper" > /opt/supervisor-scripts/my-application.sh
chmod +x /opt/supervisor-scripts/my-application.sh

# Reconfigure the instance portal
rm -f /etc/portal.yaml
export PORTAL_CONFIG="localhost:1111:11111:/:Instance Portal|localhost:1234:11234:/:My Application"

# Reload Supervisor
supervisorctl reload
```

--------------------------------

### Install and start SGLang router

Source: https://docs.vast.ai/examples/serving-infrastructure/sglang-router-vast

Creates a virtual environment, installs sglang-router, and starts the router with two worker endpoints using round-robin policy. The router listens on port 30000 and distributes requests to the specified workers.

```bash
# Create virtual environment and install router
uv venv .venv && source .venv/bin/activate
uv pip install sglang-router

# Start router with both worker endpoints
python -m sglang_router.launch_router \
    --host 0.0.0.0 \
    --port 30000 \
    --worker-urls http://<WORKER1_IP>:<WORKER1_PORT> http://<WORKER2_IP>:<WORKER2_PORT> \
    --policy round_robin
```

--------------------------------

### vastai start instances usage

Source: https://docs.vast.ai/cli/reference/start-instances

Shows the command syntax for starting instances. The IDs are required and must be provided as arguments.

```bash
vastai start instances [OPTIONS] ID0 ID1 ID2...
```

--------------------------------

### cURL GET /api/v0/endptjobs

Source: https://docs.vast.ai/api-reference/serverless/show-endpoints

cURL example for the GET /api/v0/endptjobs endpoint. Requires an Authorization header with a bearer token.

```bash
curl --request GET \
  --url https://console.vast.ai/api/v0/endptjobs \
  --header 'Authorization: Bearer <token>'
```

--------------------------------

### Example: Launch an instance

Source: https://docs.vast.ai/sdk/python/reference/launch-instance

Shows how to instantiate the VastAI client and call launch_instance with the required parameters. Replace YOUR_API_KEY with your actual API key.

```python
from vastai import VastAI

client = VastAI(api_key="YOUR_API_KEY")
result = client.launch_instance(gpu_name="value", num_gpus="value", image="value")
print(result)
```

--------------------------------

### Set up working directory and install requests

Source: https://docs.vast.ai/examples/notifications/google-chat-webhook

Creates a virtual environment and installs the only Python dependency, requests, needed to run the adapter.

```bash
mkdir vast-google-chat-notifications
cd vast-google-chat-notifications
python3 -m venv .venv
. .venv/bin/activate
pip install requests

```

--------------------------------

### Ruby GET /api/v0/endptjobs

Source: https://docs.vast.ai/api-reference/serverless/show-endpoints

Ruby example using net/http to call the GET /api/v0/endptjobs endpoint. Requires an Authorization header with a bearer token.

```ruby
require 'uri'
require 'net/http'

url = URI("https://console.vast.ai/api/v0/endptjobs")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Get.new(url)
request["Authorization"] = 'Bearer <token>'

response = http.request(request)
puts response.read_body
```

--------------------------------

### Create instance example

Source: https://docs.vast.ai/sdk/python/reference/create-instance

Minimal example of creating an instance using the VastAI client. Requires an API key and an offer ID from search offers.

```python
from vastai import VastAI

client = VastAI(api_key="YOUR_API_KEY")
result = client.create_instance(id=12345)
print(result)
```

--------------------------------

### Create instance with vastai create instance

Source: https://docs.vast.ai/cli/get-started

Rents the machine using the offer ID from a previous search. The `--image` flag specifies the Docker image, `--disk 20` sets 20 GB of storage, `--onstart-cmd` runs a command at boot, and `--ssh --direct` enables direct SSH access. Storage charges begin at creation; GPU charges begin when the instance reaches the `running` state. The `--onstart-cmd` is limited to 16KB.

```bash
vastai create instance OFFER_ID --image pytorch/pytorch:2.4.0-cuda12.4-cudnn9-runtime --disk 20 --onstart-cmd "echo hello && nvidia-smi" --ssh --direct

```

--------------------------------

### Go GET /api/v0/endptjobs

Source: https://docs.vast.ai/api-reference/serverless/show-endpoints

Go example using net/http to call the GET /api/v0/endptjobs endpoint. Requires an Authorization header with a bearer token.

```go
package main

import (
	"fmt"
	"net/http"
	"io"
)

func main() {

	url := "https://console.vast.ai/api/v0/endptjobs"

	req, _ := http.NewRequest("GET", url, nil)

	req.Header.Add("Authorization", "Bearer <token>")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}
```

--------------------------------

### PHP GET /api/v0/endptjobs

Source: https://docs.vast.ai/api-reference/serverless/show-endpoints

PHP example using cURL to call the GET /api/v0/endptjobs endpoint. Requires an Authorization header with a bearer token.

```php
<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://console.vast.ai/api/v0/endptjobs",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "GET",
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}
```

--------------------------------

### Quick start: create client and search offers

Source: https://docs.vast.ai/sdk/python/reference/vastai

Create a VastAI client with an API key, search for GPU offers, and show instances. Requires a valid API key.

```python
from vastai import VastAI

client = VastAI(api_key="YOUR_API_KEY")

# Search for GPU offers
offers = client.search_offers(query="gpu_name=RTX_3090 rentable=True")
print(offers)

# Show your instances
instances = client.show_instances()
print(instances)
```

--------------------------------

### JavaScript GET /api/v0/endptjobs

Source: https://docs.vast.ai/api-reference/serverless/show-endpoints

JavaScript example using fetch to call the GET /api/v0/endptjobs endpoint. Requires an Authorization header with a bearer token.

```javascript
const options = {method: 'GET', headers: {Authorization: 'Bearer <token>'}};

fetch('https://console.vast.ai/api/v0/endptjobs', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));
```

--------------------------------

### Example: vastai show deployments

Source: https://docs.vast.ai/cli/reference/show-deployments

Example command to display the user's current deployments.

```bash
vastai show deployments

```

--------------------------------

### Example: Prepay Command with Output

Source: https://docs.vast.ai/guides/instances/choosing/reserved-instances

Example of prepaying $50 for instance 24973511, showing the output with the prepaid duration and discount percentage.

```bash
vastai prepay instance 24973511 50
```

```text
prepaid for 0.546 months of instance 24973511 applying $50.0 credits for a discount of 3.5000000000000004%
```

--------------------------------

### vastai get endpt-logs example

Source: https://docs.vast.ai/cli/reference/get-endpt-logs

Example command to fetch logs for an endpoint group. Replace <ID> with the actual endpoint group ID.

```bash
vastai get endpt-logs <ID>

```

--------------------------------

### Example response for GET /api/v0/charges

Source: https://docs.vast.ai/api-reference/billing/show-charges

Example JSON response showing a successful charges list with count, total, next_token, and results array.

```json
{
  "success": true,
  "count": 2,
  "total": 14,
  "next_token": "eyJ2YWx1ZXMiOiB7ImlkIjogMTIzNDU2Nzh9fQ==",
  "results": [
    {
      "start": 1730419200,
      "end": 1730678400,
      "type": "instance",
      "source": "instance-12345678",
      "description": "Instance 12345678 Charges - 4 days",
      "amount": 38.421,
      "metadata": {
        "template_id": 101,
        "label": "my-training-job",
        "endpoint_id": 19392,
        "workergroup_id": 25355
      },
      "items": [
        {
          "start": 1730419200,
          "end": 1730678400,
          "type": "gpu",
          "source": null,
          "description": "96.000 hours at $0.389/hour",
          "amount": 37.344,
          "metadata": {},
          "items": []
        }
      ]
    }
  ]
}
```

--------------------------------

### Example: client.search_offers()

Source: https://docs.vast.ai/sdk/python/reference/search-offers

Minimal usage example. Creates a VastAI client with an API key and calls search_offers() with no arguments, then prints the result list.

```python
from vastai import VastAI

client = VastAI(api_key="YOUR_API_KEY")
result = client.search_offers()
print(result)
```

--------------------------------

### Java (Unirest) GET /api/v0/endptjobs

Source: https://docs.vast.ai/api-reference/serverless/show-endpoints

Java example using Unirest to call the GET /api/v0/endptjobs endpoint. Requires an Authorization header with a bearer token.

```java
HttpResponse<String> response = Unirest.get("https://console.vast.ai/api/v0/endptjobs")
  .header("Authorization", "Bearer <token>")
  .asString();
```

--------------------------------

### Go example

Source: https://docs.vast.ai/api-reference/volumes/rent-volume

Go example using net/http to PUT a volume creation request. Sends a JSON payload and prints the response body.

```go
package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://console.vast.ai/api/v0/volumes"

	payload := strings.NewReader("{\n  \"id\": 420,\n  \"size\": 15\n}")

	req, _ := http.NewRequest("PUT", url, payload)

	req.Header.Add("Authorization", "Bearer <token>")
	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}
```

--------------------------------

### Go example

Source: https://docs.vast.ai/api-reference/instances/create-instance

Uses the net/http package to send a PUT request to create an instance. Prints the response body.

```go
package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://console.vast.ai/api/v0/asks/{id}"

	payload := strings.NewReader("{\n  \"image\": \"vastai/base-image:@vastai-automatic-tag\"\n}")

	req, _ := http.NewRequest("PUT", url, payload)

	req.Header.Add("Authorization", "Bearer <token>")
	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

```

--------------------------------

### Python GET /api/v0/endptjobs

Source: https://docs.vast.ai/api-reference/serverless/show-endpoints

Python example using the requests library to call the GET /api/v0/endptjobs endpoint. Requires an Authorization header with a bearer token.

```python
import requests

url = "https://console.vast.ai/api/v0/endptjobs"

headers = {"Authorization": "Bearer <token>"}

response = requests.get(url, headers=headers)

print(response.text)
```

--------------------------------

### End-to-End Example (Python)

Source: https://docs.vast.ai/api-reference/creating-instances-with-api

Uses the requests library to perform the same workflow: search offers, create instance, and check status. Requires an API key set as 'your_api_key'. Includes a 5-second sleep before checking status.

```python
import requests
import time

api_key = "your_api_key"
headers = {
    "Authorization": f"Bearer {api_key}",
    "Content-Type": "application/json"
}

# Step 1: Search for offers
response = requests.post(
    "https://console.vast.ai/api/v0/bundles/",
    headers=headers,
    json={
        "gpu_name": {"in": ["RTX 4090"]},
        "num_gpus": {"gte": 1},
        "gpu_ram": {"gte": 24000},
        "reliability": {"gte": 0.99},
        "verified": {"eq": True},
        "rentable": {"eq": True},
        "type": "ondemand",
        "limit": 3
    }
)

offers = response.json().get("offers", [])
for offer in offers:
    print(f"Offer {offer['id']}: {offer['num_gpus']}x {offer['gpu_name']} "
          f"- ${offer['dph_total']:.3f}/hr")

# Step 2: Create instance using the first offer
offer_id = offers[0]["id"]

response = requests.put(
    f"https://console.vast.ai/api/v0/asks/{offer_id}/",
    headers=headers,
    json={
        "image": "vllm/vllm-openai:latest",
        "label": "my-vllm-server",
        "disk": 50,
        "runtype": "ssh_direct",
        "env": {"MODEL_ID": "deepseek-ai/DeepSeek-R1-Distill-Llama-8B", "-p 8000:8000": "1"},
        "onstart": "vllm serve $MODEL_ID --port 8000"
    }
)

result = response.json()
instance_id = result.get("new_contract")
print(f"\nCreated instance: {instance_id}")

# Step 3: Wait and check instance status
time.sleep(5)

response = requests.get(
    f"https://console.vast.ai/api/v0/instances/{instance_id}/",
    headers=headers
)

instance = response.json().get("instances", {})
print(f"Status: {instance.get('actual_status')}")
print(f"Label: {instance.get('label')}")

```

--------------------------------

### Java Unirest GET /api/v0/secrets

Source: https://docs.vast.ai/api-reference/accounts/show-env-vars

Java example using Unirest to make a GET request to the /api/v0/secrets endpoint. Replace <token> with your API token.

```java
HttpResponse<String> response = Unirest.get("https://console.vast.ai/api/v0/secrets")
  .header("Authorization", "Bearer <token>")
  .asString();
```

--------------------------------

### End-to-End Example (curl)

Source: https://docs.vast.ai/api-reference/creating-instances-with-api

Uses curl and jq to search for RTX 4090 offers, create an instance with vLLM, and check its status. Requires VAST_API_KEY environment variable. The onstart command runs vLLM serve on port 8000.

```bash
# Step 1: Search for offers
OFFERS=$(curl -s -X POST "https://console.vast.ai/api/v0/bundles/" \
  -H "Authorization: Bearer $VAST_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "gpu_name": {"in": ["RTX 4090"]},
    "num_gpus": {"gte": 1},
    "gpu_ram": {"gte": 24000},
    "reliability": {"gte": 0.99},
    "verified": {"eq": true},
    "rentable": {"eq": true},
    "type": "ondemand",
    "limit": 3
  }')

echo "Available offers:"
echo "$OFFERS" | jq '.offers[] | {id, gpu_name, num_gpus, dph_total}'

# Step 2: Create instance using the first offer
OFFER_ID=$(echo "$OFFERS" | jq '.offers[0].id')

RESULT=$(curl -s -X PUT "https://console.vast.ai/api/v0/asks/$OFFER_ID/" \
  -H "Authorization: Bearer $VAST_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "image": "vllm/vllm-openai:latest",
    "label": "my-vllm-server",
    "disk": 50,
    "runtype": "ssh_direct",
    "env": {"MODEL_ID": "deepseek-ai/DeepSeek-R1-Distill-Llama-8B", "-p 8000:8000": "1"},
    "onstart": "vllm serve $MODEL_ID --port 8000"
  }')

INSTANCE_ID=$(echo "$RESULT" | jq '.new_contract')
echo "Created instance: $INSTANCE_ID"

# Step 3: Check instance status
curl -s "https://console.vast.ai/api/v0/instances/$INSTANCE_ID/" \
  -H "Authorization: Bearer $VAST_API_KEY" | jq '{id: .instances.id, status: .instances.actual_status, label: .instances.label}'

```

--------------------------------

### PHP cURL GET /api/v0/secrets

Source: https://docs.vast.ai/api-reference/accounts/show-env-vars

PHP example using cURL to make a GET request to the /api/v0/secrets endpoint. Replace <token> with your API token.

```php
<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://console.vast.ai/api/v0/secrets",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "GET",
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}
```

--------------------------------

### Example usage of VastAI.show_deployments()

Source: https://docs.vast.ai/sdk/python/reference/show-deployments

Creates a VastAI client with an API key, calls show_deployments(), and prints the result. Requires a valid API key.

```python
from vastai import VastAI

client = VastAI(api_key="YOUR_API_KEY")
result = client.show_deployments()
print(result)

```

--------------------------------

### Install client dependencies with pip

Source: https://docs.vast.ai/examples/ocr/rolmocr

Install the required Python packages: openai, datasets, and pydantic. Use this command to set up the environment before running the OCR example.

```bash
pip install --upgrade openai datasets pydantic

```

--------------------------------

### Get instance ID from VAST_CONTAINERLABEL

Source: https://docs.vast.ai/guides/reference/faq/instances

Use the VAST_CONTAINERLABEL environment variable to get the instance ID from within the container. The output shows an example ID C.38250.

```bash
echo $VAST_CONTAINERLABEL
# Output: C.38250
```

--------------------------------

### vastai get wrkgrp-logs example

Source: https://docs.vast.ai/cli/reference/get-wrkgrp-logs

Example command to fetch logs for a worker group, using a placeholder ID. The ID is required and must be replaced with the actual endpoint group ID.

```bash
vastai get wrkgrp-logs <ID>

```

--------------------------------

### Start instance with explicit API key

Source: https://docs.vast.ai/guides/instances/docker-environment

Starts the instance while explicitly providing the per-instance API key. Use this if the API key is not automatically configured.

```bash
vastai start instance $CONTAINER_ID --api-key $CONTAINER_API_KEY

```

--------------------------------

### Search offers and create instance with Vast CLI

Source: https://docs.vast.ai/examples/migrations/runpod-to-vast

Two-step workflow: search for matching offers using a filter query, then create an instance from an offer ID. The search filters by GPU name and reliability, orders by base price, and limits to 5 results.

```bash
# Step 1: Search for matching offers
vastai search offers 'gpu_name = RTX_4090 reliability > 0.95' \
    --order "dph_base" --limit 5

# Step 2: Create instance from an offer ID
vastai create instance <OFFER_ID> \
    --image pytorch/pytorch:latest \
    --disk 50

```

--------------------------------

### PHP: GET /api/v0/template

Source: https://docs.vast.ai/api-reference/search/search-templates

PHP cURL example to fetch templates. Sets the Authorization header and handles errors.

```php
<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://console.vast.ai/api/v0/template",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "GET",
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}
```

--------------------------------

### JavaScript: GET /api/v0/template

Source: https://docs.vast.ai/api-reference/search/search-templates

JavaScript fetch example to retrieve templates. Uses the Authorization header with a bearer token.

```javascript
const options = {method: 'GET', headers: {Authorization: 'Bearer <token>'}};

fetch('https://console.vast.ai/api/v0/template', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));
```

--------------------------------

### Go: Start deployment

Source: https://docs.vast.ai/api-reference/serverless/start-deployment

Uses net/http to POST to the start deployment endpoint. Prints the response body.

```go
package main

import (
	"fmt"
	"net/http"
	"io"
)

func main() {

	url := "https://console.vast.ai/api/v0/deployment/{id}/start"

	req, _ := http.NewRequest("POST", url, nil)

	req.Header.Add("Authorization", "Bearer <token>")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}
```

--------------------------------

### Create an instance with vast.create_instance() and view output

Source: https://docs.vast.ai/sdk/python/quickstart

Rents a machine using create_instance with the offer ID from the search step. The output includes the new instance ID (new_contract). Storage charges begin at creation; GPU charges begin when the instance reaches the running state.

```python
result = vast.create_instance(
    id=OFFER_ID,
    image="pytorch/pytorch:2.4.0-cuda12.4-cudnn9-runtime",
    disk=20,
    onstart_cmd="echo hello && nvidia-smi",
    ssh=True,
    direct=True,
)
print(result)

```

```json
{"success": true, "new_contract": 12345678}

```

--------------------------------

### Create instance with on-start script via Vast CLI

Source: https://docs.vast.ai/examples/migrations/salad-to-vast

Vast CLI command to create an instance with an on-start script using --onstart-cmd. The script runs after the container starts, useful for installing dependencies or starting services.

```bash
vastai create instance <OFFER_ID> \
    --image pytorch/pytorch:latest \
    --env "-p 8000:8000" \
    --disk 50 \
    --onstart-cmd "cd /workspace && pip install -r requirements.txt && python server.py"

```

--------------------------------

### vastai start instance usage

Source: https://docs.vast.ai/cli/reference/start-instance

Shows the command syntax for starting an instance. The ID argument is required and specifies the instance to start or restart.

```bash
vastai start instance ID [OPTIONS]

```

--------------------------------

### Example usage of VastAI.show_maints

Source: https://docs.vast.ai/host/sdk/show-maints

Example showing how to instantiate the VastAI client and call show_maints with a machine ID string, then print the result.

```python
from vastai import VastAI

client = VastAI(api_key="YOUR_API_KEY")
result = client.show_maints(ids="...")
print(result)

```

--------------------------------

### Success response for GET notification preferences

Source: https://docs.vast.ai/api-reference/notifications/get-notification-preferences

Example 200 response showing the notification_preferences object with a low_credit client preference.

```json
{
  "success": true,
  "notification_preferences": {
    "client": {
      "low_credit": {
        "email": true,
        "webhooks": false
      }
    }
  }
}
```

--------------------------------

### vastai update template example with options

Source: https://docs.vast.ai/cli/reference/update-template

Example that updates a template with a new name, image, environment variables, onstart command, search parameters, disk space, and SSH/direct flags.

```bash
vastai update template c81e7ab0e928a508510d1979346de10d --name "tgi-llama2-7B-quantized" --image "ghcr.io/huggingface/text-generation-inference:1.0.3"
                            --env "-p 3000:3000 -e MODEL_ARGS='--model-id TheBloke/Llama-2-7B-chat-GPTQ --quantize gptq'"
                            --onstart-cmd 'wget -O - https://raw.githubusercontent.com/vast-ai/vast-pyworker/main/scripts/launch_tgi.sh | bash'
                            --search_params "gpu_ram>=23 num_gpus=1 gpu_name=RTX_3090 inet_down>128 direct_port_count>3 disk_space>=192 driver_version>=535086005 rented=False"
                            --disk 8.0 --ssh --direct
```

--------------------------------

### Example output of systemctl status vastai

Source: https://docs.vast.ai/host/machine-offline

Example output when the service is running normally. The 'Active: active (running)' line and the 'since' timestamp indicate stability. A very recent start time or 'activating (auto-restart)' state signals a restart loop.

```text
● vastai.service - Vast.ai Host Daemon
     Loaded: loaded (/etc/systemd/system/vastai.service; enabled; preset: enabl>
     Active: active (running) since Tue 2026-08-04 07:47:51 +07; 6h ago
   Main PID: 2966 (launch_kaalia.s)
      Tasks: 2 (limit: 143635)
     Memory: 3.5G (peak: 6.5G)
        CPU: 5min 41.404s
     CGroup: /system.slice/vastai.service
             ├─2966 /bin/bash /var/lib/vastai_kaalia/latest/launch_kaalia.sh
             └─2975 /var/lib/vastai_kaalia/latest/kaalia backend=DKR installpat>

```

--------------------------------

### PHP cURL GET for show deposit

Source: https://docs.vast.ai/api-reference/billing/show-deposit

PHP cURL example for the show deposit endpoint. Replace {id} and <token> as needed.

```php
<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://console.vast.ai/api/v0/instances/balance/{id}",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "GET",
  CURLOPT_HTTPHEADER => [
    "Authorization: Bearer <token>"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}
```

--------------------------------

### Import Vast.ai plugin in Cursor

Source: https://docs.vast.ai/guides/get-started/agents

Install from the Cursor Dashboard → Settings → Plugins → Team Marketplaces → Import, then paste the URL. Then say “Configure my Vast.ai CLI.” in Agent chat to walk through setup.

```text
https://github.com/vast-ai/vast-cursor-plugin
```

--------------------------------

### Go Example

Source: https://docs.vast.ai/api-reference/accounts/create-api-key

Go example using net/http to POST to the API endpoint. Prints the response body.

```go
package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://console.vast.ai/api/v0/auth/apikeys"

	payload := strings.NewReader("{\n  \"name\": \"read-only-key\",\n  \"permissions\": {\n    \"read\": true,\n    \"write\": false\n  },\n  \"key_params\": {\n    \"ip_whitelist\": [\n      \"1.2.3.4\"\n    ]\n  }\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Authorization", "Bearer <token>")
	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}
```

--------------------------------

### JavaScript fetch GET for show deposit

Source: https://docs.vast.ai/api-reference/billing/show-deposit

JavaScript fetch example for the show deposit endpoint. Replace {id} and <token> as needed.

```javascript
const options = {method: 'GET', headers: {Authorization: 'Bearer <token>'}};

fetch('https://console.vast.ai/api/v0/instances/balance/{id}', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));
```

--------------------------------

### Java (Unirest): GET /api/v0/template

Source: https://docs.vast.ai/api-reference/search/search-templates

Java example using Unirest to fetch templates. Sets the Authorization header and returns the response as a string.

```java
HttpResponse<String> response = Unirest.get("https://console.vast.ai/api/v0/template")
  .header("Authorization", "Bearer <token>")
  .asString();
```

--------------------------------

### Example usage of show_deployment

Source: https://docs.vast.ai/sdk/python/reference/show-deployment

Example of creating a VastAI client and calling show_deployment with an id, then printing the result.

```python
from vastai import VastAI

client = VastAI(api_key="YOUR_API_KEY")
result = client.show_deployment(id=12345)
print(result)

```
Tokens:
Raw