DOC · MICRODC-PROV-001v3.4 · BETA

Connect your hardware to MicroDC.ai and start earning while you sleep. Automatic job distribution, transparent per-job ledger, payouts your way. Multi-GPU and concurrent jobs supported. CPU-only welcome.

Maximize your hardware investment.

§01 · WHY PROVIDE
CAP · 01

5-minute setup

Generate a worker token, install, start. Jobs flow automatically.

CAP · 02

Automatic jobs

Scheduler matches jobs to your hardware. No manual selection.

CAP · 03

Transparent payouts

Per-job ledger. Track earnings in real time on your dashboard.

CAP · 04

Full control

Choose models, set capacity, manage availability — start and stop anytime.

Three steps to start earning.

§02 · ONBOARDING
STEP 01━━━▶

Create account

Sign up for a free MicroDC.ai account and generate a worker registration token in the console.

STEP 02━━━▶

Install worker

Download and configure the worker software on your GPU machine. Takes about 5 minutes.

STEP 03

Start earning

Your worker automatically picks up jobs from the queue. Watch earnings accumulate in your dashboard.

Hardware requirements.

§03 · SPECS
GPU · RECOMMENDED
  • NVIDIA GPU with 8GB+ VRAM (RTX 3080, 4080, A100, ...)
  • 16GB+ system RAM
  • Stable internet (50+ Mbps recommended)
  • Linux or Windows with Docker support
CPU · ALSO SUPPORTED
  • Modern multi-core CPU (8+ cores recommended)
  • 32GB+ system RAM for larger models
  • Suitable for embedding models and smaller LLMs
  • Great for getting started without a GPU

Earnings & payouts, your way.

§04 · PAYOUTS
CREDITS · WITH BONUS

Instant platform credits

Use them on your own jobs or save them. A bonus percentage is added on every payout.

  • Instant — no processing delay
  • Bonus % on every credit payout
  • Great for hybrid users (provider + customer)
PAYPAL · CASH

Real USD payout

Request to your PayPal account. Submitted for processing and recorded as a PAYOUT transaction in your billing history.

  • Real USD, direct to your PayPal
  • Full ledger trail for taxes and accounting
  • No minimum balance quirks
ALL-TIME
Lifetime
Earnings
ROLLING
30/90 day
Time windows
BREAKDOWN
Per-worker
Sortable table
AUDIT
Full
Payout history

Organize fleets with worker groups.

§05 · FLEET MGMT

Running more than one machine? Group them — by site, by GPU tier, by customer. Aggregate earnings stats per group, sourced from the real job-assignment ledger.

CAP · 01

Name your groups

"Home-Lab", "Colo-Rack-A", "Customer-X". Any name, per-user unique.

CAP · 02

Aggregate stats

Earnings totals per group over 30-day, 90-day, and all-time.

CAP · 03

Manage inline

Add or remove workers right from the group detail panel. Delete a group — workers stay put.

Low-spec hardware? You're welcome here.

§06 · CONTEXT TIERS

Your worker declares its prompt-size capacity via heartbeat. The scheduler sends you jobs at or below that tier — nothing bigger. An Orange Pi Max 16GB serves short prompts fast; an H100 gets the 24K-char monsters. Everybody wins.

TierPrompt sizeHardware fit
Tier 10–1.5K charsLow-end GPU / CPU, ARM SBC, Orange Pi
Tier 21.5K–6K charsEntry-level GPU, 8GB VRAM
Tier 36K–24K charsMid-range GPU, 16GB+ VRAM
Tier 424K+ charsHigh-end GPU, 24GB+ VRAM

DON'T SET A TIER? NO PROBLEM — EXISTING WORKERS KEEP SEEING ALL JOBS. TIERS ARE OPT-IN.

Opt into more job types.

§07 · CAPABILITIES
CAP · DOCKER

Container jobs

Advertise capabilities: ["docker"] to receive customer container jobs. GPU-hour or CPU-core-hour pricing.

CAP · ENCRYPTION

Zero-knowledge

Advertise capabilities: ["encryption"] to handle end-to-end encrypted workloads.

CAP · CONCURRENT

Multi-job

Set max_concurrent_jobs > 1 to claim multiple jobs at once. Marked BUSY only at full capacity.

CAP · CPU-ONLY

No GPU? Fine.

Workers without GPUs heartbeat cleanly and serve embedding / small-LLM jobs at tier 1–2.

§08 · ENGAGE

Plug in. Earn.

Free to register. Five-minute install. First payout when you're ready.