A distributed load-testing platform built 100% by AI agents.
Turn API & business flows into reusable load tests.
A self-hosted distributed API load-testing platform—and an inspectable experiment in AI-agent software delivery.
How it was built.
From intent to running platform — orchestrated by AI agents.
Product Intent
Product intent and requirements discussed with a human through text.
AI Product Design
AI agents define product design, governance, quality gates, and specifications.
AI Architecture
AI agents define architecture, contracts, and the interface system.
AI Implementation
AI agents author application code, tests, and verification repairs.
AI Verification & Release
AI agents execute quality gates and produce deployment and release assets.
The original baseline — produced by AI agents.
Human participation is explicit: product intent, requirements discussion, product use, usage feedback, and result acceptance. AI agents handled product and process design, governance, quality gates, specifications, architecture, implementation, tests, verification repair, deployment, and release assets.
design → cross-review → code → review ⇄ fix → frozen ↺ re-iterate
Command every request.
Scalable infrastructure that stays within your private network, giving you total control over data and execution.
Visual Scenario Designer
Design complex API flows visually. Capture requests, define headers, and set assertions without writing complex code.
Test Plan Orchestration
Chain multiple scenarios together. Control ramp-up, hold times, and failure criteria at a global level.
Scalable Generator Mesh
Deploy worker nodes as a distributed cluster. Orchestrate massive load from your own private infrastructure.
Traceable Run Reports
Trace every test run. Detailed interactive reports including performance artifacts, logs, and failure analysis.
Unified Asset Management
Centralized storage for datasets, external scripts, and configuration files. Upload once, reuse across all plans.
Open Source & Privacy First
No proprietary lock-in. Host it on-premise or in your cloud. Built for privacy-conscious engineering teams.
Recent Test Runs
UI DEMONSTRATION · EXAMPLE DATA · NOT A LIVE SERVICE · NOT BENCHMARK DATA| RUN ID | STATUS | SLA RESULT | P95 LATENCY | ERROR % | DURATION |
|---|---|---|---|---|---|
| #SP-9482 | FINISHED | PASSED | 142ms | 0.00% | 15m 00s |
| #SP-9481 | RUNNING | PENDING | 218ms | 0.84% | 08m 42s |
| #SP-9480 | FAILED | Critical | 1,240ms | 22.4% | 02m 14s |
| #SP-9479 | ABORTED | N/A | -- | 0.00% | 00m 18s |
| #SP-9478 | FINISHED | PASSED | 94ms | 0.01% | 10m 00s |
| #SP-9477 | FINISHED | PASSED | 108ms | 0.00% | 12m 30s |
| #SP-9476 | FAILED | Soft Fail | 356ms | 1.20% | 05m 00s |
| #SP-9475 | FINISHED | PASSED | 112ms | 0.00% | 15m 00s |
Distributed load mesh
High-scale orchestration across distributed clusters with complete visibility.
UI DEMONSTRATION · EXAMPLE DATA · NOT A LIVE SERVICE · NOT BENCHMARK DATA
Ready in four steps.
From concept to production-ready performance reports.
1. Design
Design a performance scenario visually or through configuration.
2. Orchestrate
Combine scenarios into a robust, high-scale Test Plan.
3. Run
Execute at scale from your distributed Load Nodes.
4. Report
Analyze deep-dive metrics with traceable run reports.
Inspect the baseline.
Then judge the claim.
Open Source • MIT License • Self-Hosted