CLUSTER: ONLINE (Enterprise High-VRAM GPU Cluster)
P50 TTFT: 28ms KV-CACHE: 78.4% 0% US-CLOUD
dp
dev.pipe Sovereign Bare-Metal Inference
[CORE: DEVPIPE-CODE-1] // DEDICATED BARE-METAL SILICON FLATRATE

High-Throughput Inference for your Code.
True Flatrate. 0 % US Cloud.

Eliminate token anxiety and transatlantic data transfers. Dedicated Bare Metal in Germany [GER-01]. True Flatrate for Cursor, Cline, Continue, and JetBrains with Zero-Retention VRAM Inference.

10-SECOND IDE INSTALLER (BASH / POWERSHELL) NO CREDIT CARD REQUIRED
$ curl -fsSL https://www.devpipe.eu/try | bash

Automatically detects Cursor, Continue.dev, Cline & VS Code and configures OpenAI-compatible endpoints.

PRICING 29 € / Month Flat incl. VAT
LOCATION Germany (EU)
GDPR ART. 28 1-Click DPA Included

LIVE TELEMETRY (FRA-01)

Real-time diagnostics of the Enterprise High-VRAM GPU Cluster

LIVE
SINGLE-STREAM SPEED
112 Tok/s
P50 FIRST-TOKEN LATENCY
28 ms
RADIX PREFIX CACHE
78.4 %
GPU QUEUE DEPTH
0 Queued
UNLIMITED VOLUME, VARIABLE SPEED RADIX PREFIX CACHE

Repeated agent context hits the RadixAttention prefix cache instead of being recomputed — that is what makes a flatrate without volume caps affordable to run. Paid plans are never metered by volume; under peak load only the throughput per key is adjusted.

// THE POSITIONING

The Uncapped Agent Workhorse Tier.

We don't claim to replace expensive frontier models for complex multi-system architecture design.
Save your Claude tokens for the architecture. Offload your 10,000 background loops to dev.pipe.

US CLOUD PAY-PER-TOKEN Metered API
200 € – 600 € / month
  • Panic over infinite Composer & Tool agent loops
  • Transatlantic data transfer (US CLOUD Act risk)
  • Unpredictable end-of-month API invoice spikes
dev.pipe WORKHORSE FLATRATE SOVEREIGN FLAT
29 € / month incl. VAT
  • Uncapped refactoring, linting, test generation & agent loops
  • 100% German Bare-Metal Infrastructure
  • Save expensive Claude 3.5 Sonnet tokens for high-level architecture

The “Token-Burn” ROI Calculator

Agents like Cline or Cursor Composer send full repository context with every refactoring prompt. Calculate your real savings compared to pay-per-token cloud APIs.

25 Refactorings / Day
5 Prompts (Casual Dev) 40 Prompts (Power-User) 80 Prompts (Full-Agentic Team)
MONTHLY VOLUME
~19.3 Tokens / Month
US CLOUD COST (PAY/TOK)
~106 €
dev.pipe FLATRATE
29 €
YOUR SAVINGS
77 € / Month (73% Savings)
Get Flatrate Now →

dev.pipe flatrate shown incl. VAT. Cloud costs are estimated from net list prices before VAT.

30-Second Drop-In Integration

dev.pipe provides a standard OpenAI-compatible interface. Works seamlessly with all major editors and coding extensions.

In Cursor, navigate to Settings → Models → OpenAI API Key and enable Override OpenAI Base URL:

OpenAI Base URL: https://api.devpipe.eu/v1
OpenAI API Key: sk-live-your-devpipe-key...
Model Name: devpipe-code-1

True Flatrate. Absolute Cost Predictability.

No hidden fees, no credit card billing surprises. Every plan includes full GDPR legal compliance with an Art. 28 Data Processing Agreement (DPA).

Solo Flatrate

RECOMMENDED
29 € / month incl. VAT

For freelancers & senior engineers with always-on coding agents.

  • Unlimited token volume — the flatrate has no monthly cap
  • Variable speed: 60,000 TPM baseline per key
  • 1 API key
  • 1-Click digital DPA (GDPR Art. 28)

Team Flatrate

5 DEVELOPERS
249 € / month incl. VAT

For developer teams, agencies, and companies with strict NDAs.

  • Unlimited token volume — the flatrate has no monthly cap
  • Variable speed: 450,000 TPM pooled across the whole team
  • Idle seats lend their speed: when only two of five developers are heads-down, they share the full team throughput — five separate Solo plans cannot do that
  • 5 seats — one API key per developer
  • One Art. 28 DPA in your company's name — employee-held Solo plans give the company no contract of its own
  • Itemized invoice with VAT identification