Quick Start

Overview

Get your first inference job running in 5 minutes. This guide covers the developer path — SDK installation, API key setup, and your first API call.

**Prerequisites:** Node.js 18+ or Python 3.9+

---

Step 1: Install the SDK

Pick your language:

# Node.js / TypeScript
npm install dc1-renter-sdk

# Python
pip install dc1

No other dependencies required. Both SDKs are thin wrappers around the REST API.

---

Step 2: Get Your API Key

1. Sign up at https://dcp.sa/renter/register
2. Go to **Dashboard → Settings → API Keys**
3. Copy your key — it starts with `dc1-renter-`

Store it securely:

# Node.js
export DC1_RENTER_KEY="dc1-renter-abc123..."

# Python
export DC1_API_KEY="dc1-renter-abc123..."

---

Step 3: Verify Your Key Works

Node.js

import { DC1RenterClient } from 'dc1-renter-sdk';

const client = new DC1RenterClient({ apiKey: process.env.DC1_RENTER_KEY! });
const me = await client.me();
console.log(`Balance: ${me.balanceSar} SAR`);

Python

import dc1

client = dc1.DC1Client(api_key=os.environ["DC1_API_KEY"])
wallet = client.wallet.balance()
print(f"Balance: {wallet.balance_sar} SAR")

Expected output: `Balance: 50.00 SAR` (new accounts get 50 SAR test credit).

---

Step 4: List Available GPUs

Node.js

const providers = await client.listProviders();
const gpus = providers.filter(p => p.vramGb >= 24);
console.log(`Found ${gpus.length} GPUs with 24+ GB VRAM`);

for (const p of gpus.slice(0, 3)) {
  console.log(`  ${p.name} | ${p.gpuModel} | ${p.vramGb} GB | ${p.reliabilityScore}% reliability`);
}

Python

providers = client.providers.list()
gpus = [p for p in providers if p.vram_gb >= ]
print(f"Found {len(gpus)} GPUs with + GB VRAM")

for p in gpus[:]:
    print(f"  {p.name} | {p.gpu_model} | {p.vram_gb} GB | {p.reliability_score}% reliability")

cURL

curl -s "https://dcp.sa/api/dc1/renters/available-providers" | \
  jq &#;.providers[:] | .[] | {name, gpu_model, vram_gb, reliability_score}&#;

---

Step 5: Submit Your First Job

Node.js

const job = await client.submitJob({
  providerId: providers[0].id,
  jobType: 'llm_inference',
  params: {
    prompt: 'Explain the Vision 2030 initiative in 2 sentences.',
    model: 'TinyLlama/TinyLlama-1.1B-Chat-v1.0',
  },
  durationMinutes: 2,
});

console.log(`Job submitted: ${job.id}`);

const result = await client.waitForJob(job.id, { timeout: 120_000 });
console.log(`Result: ${result.result?.output}`);

Python

job = client.jobs.submit(
    &#;llm_inference&#;,
    {&#;prompt&#;: &#;Explain the Vision  initiative in  sentences.&#;, &#;model&#;: &#;TinyLlama/TinyLlama-.1B-Chat-v1.&#;},
    provider_id=providers[].id,
    duration_minutes=,
)

print(f"Job submitted: {job.id}")

result = client.jobs.wait(job.id, timeout=)
print(f"Result: {result.result[&#;output&#;]}")

cURL

# Submit job
JOB_RESPONSE=$(curl -s -X POST "https://dcp.sa/api/dc1/jobs/submit" \
  -H "Content-Type: application/json" \
  -H "x-renter-key: $DC1_RENTER_KEY" \
  -d &#;{
    "provider_id": ,
    "job_type": "llm_inference",
    "params": {
      "prompt": "Explain the Vision  initiative in  sentences.",
      "model": "TinyLlama/TinyLlama-.1B-Chat-v1."
    },
    "duration_minutes": 
  }&#;)

JOB_ID=$(echo $JOB_RESPONSE | jq -r &#;.job.job_id&#;)
echo "Job ID: $JOB_ID"

# Poll until done (check every 5s)
while true; do
  STATUS=$(curl -s "https://dcp.sa/api/dc1/jobs/$JOB_ID" -H "x-renter-key: $DC1_RENTER_KEY" | jq -r &#;.job.status&#;)
  echo "Status: $STATUS"
  if [ "$STATUS" = "completed" ] || [ "$STATUS" = "failed" ]; then
    break
  fi
  sleep 
done

# Get result
curl -s "https://dcp.sa/api/dc1/jobs/$JOB_ID/output" -H "x-renter-key: $DC1_RENTER_KEY" | jq &#;.&#;

---

Available Models

Use `TinyLlama` for testing (cheapest, fastest). Switch to larger models for production.

---

SDK Configuration Reference

Node.js `DC1RenterClient`

const client = new DC1RenterClient({
  apiKey: 'dc1-renter-abc123',    // required
  baseUrl: 'https://api.dcp.sa',  // optional, default
  timeoutMs: 30_000,               // optional, default 30s
});

Python `DC1Client`

client = dc1.DC1Client(
    api_key=&#;dc1-renter-abc123&#;,  # required
    base_url=&#;https://api.dcp.sa&#;, # optional
    timeout=,                    # optional, seconds
)

---

Troubleshooting

`AuthError: Invalid API key`

Your key is missing or malformed.

# Verify your key format (should be dc1-renter-xxxxx)
echo $DC1_RENTER_KEY

If invalid:

1. Go to https://dcp.sa/renter/dashboard → Settings → API Keys
2. Delete the old key and create a new one
3. Copy the new key and update your environment variable

---

`APIError: 402 Payment Required`

Insufficient balance. New accounts get 50 SAR free credit.

# Check balance via API
curl -s "https://dcp.sa/api/dc1/renters/me?key=$DC1_RENTER_KEY" | jq &#;.renter.balance_halala&#;

If balance is 0:

• Add credit at https://dcp.sa/renter/billing
• Or wait for promotional credit

---

`JobTimeoutError: Job still running after N ms`

The job took longer than the timeout. Common causes:

• Provider went offline mid-job (auto-refund issued)
• Model loading took too long on first request
• Network issues

// Increase timeout for slow models
const result = await client.waitForJob(job.id, { timeout: 300_000 }); // 5 min

For persistent timeouts, try a faster model (TinyLlama) or a different provider.

---

`APIError: 400 Bad Request`

Check your request body matches the expected schema:

# Verify job submission fields
curl -s -X POST "https://dcp.sa/api/dc1/jobs/submit" \
  -H "x-renter-key: $DC1_RENTER_KEY" \
  -H "Content-Type: application/json" \
  -d &#;{"provider_id": , "job_type": "llm_inference", "params": {}, "duration_minutes": }&#; | jq &#;.error&#;

Common causes:

• Missing required `params` fields (e.g., `prompt` for `llm_inference`)
• `provider_id` doesn't exist or is offline
• `duration_minutes` is 0 or negative

---

Job Failed — "Provider Offline"

The GPU provider went offline during job execution. You receive a full automatic refund.

# Check provider status
curl -s "https://dcp.sa/api/dc1/providers/{provider_id}?key=$DC1_RENTER_KEY" | jq &#;.provider.status&#;

Resubmit your job — 99.5% of jobs complete successfully on retry.

---

Rate Limiting (429)

Too many requests. Back off and retry:

// SDK handles retries automatically, but you can add manual backoff
async function withRetry(fn, maxRetries = 3) {
  for (let i = 0; i < maxRetries; i++) {
    try {
      return await fn();
    } catch (e) {
      if (e.statusCode === 429 && i < maxRetries - 1) {
        await new Promise(r => setTimeout(r, 1000 * (i + 1)));
        continue;
      }
      throw e;
    }
  }
}

---

Python: `ModuleNotFoundError: No module named 'dc1'`

# Verify installation
pip show dc1

# Reinstall if missing
pip install dc1

Make sure you're using the correct Python environment (virtualenv, conda, etc.).

---

Node.js: `ERR_REQUIRE_ESM`

Use ESM imports or set `"type": "module"` in your `package.json`:

{
  "type": "module",
  "dependencies": {
    "dc1-renter-sdk": "^.0."
  }
}

---

Next Steps

• **API Reference** — Full endpoint documentation
• **SDK Reference (Node.js)** — SDK docs
• **SDK Reference (Python)** — SDK docs
• **Example Scripts** — Runnable examples

**Need help?** Email support@dcp.sa