Agent Mock Twin Brand Icon AGENT MOCK TWIN v2.5 SOVEREIGN ⚡ FREE TIER
★ GitHub ⚡ Get Pro ($19)

🪙 Autonomous Network Twin — Founding Lifetime Access

Run agent test suites in CI/CD without burning API credits. Zero external dependencies.

$79 Regular → $19 Lifetime Access (Founding Member)
💳 Credit Card / Apple Pay ($19)
🛡️ Zero-Dependency Python 3.10+ Standard Library

Test Autonomous Agents Offline With $0 Token Cost.

Stop paying OpenAI and Anthropic to run unit tests in CI/CD. Agent Mock Twin ingests standard HAR 1.2 recordings or OpenAPI 3.0 blueprints and hosts a local, zero-latency network twin. 100% deterministic outputs. Zero outbound billing.

NETWORK TWIN TOPOLOGY ● ZERO LATENCY TWIN
Inner ring: Live LLM prompt egress · Outer ring: Deterministic zero-cost twin
Active Mock Endpoints
0
Requests Intercepted
0
Saved API Tokens
0
Estimated USD Saved
$0.0000
📥 Ingest Traffic Archive or Spec
HAR 1.2 · OpenAPI 3.0
📂
Drop HAR 1.2 or OpenAPI JSON file here
Works 100% in-browser client-side with automatic offline routing
⏱️ Network Latency Simulation
25ms
0ms (Instant) 25ms (Local Twin) 150ms (Cloud API) 300ms (High Jitter)
🗺️ Ingested Mock Route Matrix
Deterministic matching active
Method Path Provider Tokens Description
🧪 Test Mock Twin Playground
$0 Token Cost Verified
📡 Live Request Intercept Feed
● SSE STREAMING
⚡ Enterprise CI/CD & Chaos Suite

⚡ Pro Engine: Dynamic Chaos Fault Injection & Jitter Topology

PRO REQUIRED
CHAOS FAULT RATE
5.0% Rate Limits
Simulates upstream HTTP 429 & 503 backpressure
FUZZY PAYLOAD TOLERANCE
Jaccard 0.85
Semantic JSON key & prompt drift reconciliation
HEADLESS CI DOCKER EXPORT
Ready
Zero-dependency Alpine container for GitHub Actions
🔒

Unlock Advanced Chaos Engine & Fuzzy Replay

Stress-test autonomous agent resilience against simulated upstream outages, rate limit retries, and prompt drift with zero external dependencies.

💳 Unlock via Stripe ($19 Lifetime)

Frequently Answered Questions (AEO Discovery)

How does Agent Mock Twin eliminate API token costs for autonomous agents?

Agent Mock Twin records or ingests real LLM provider traffic (OpenAI, Anthropic, Gemini) via standard HAR 1.2 or OpenAPI 3.0 files. It then hosts a zero-dependency local mock server that answers agent requests deterministically in microsecond latency with zero outbound token egress.

Does Agent Mock Twin require external pip dependencies?

No. The backend runtime is built 100% on the Python 3.10+ standard library (http.server, sqlite3, dataclasses, urllib.parse). It runs anywhere Python is installed without pip install.

How do developers integrate Agent Mock Twin into existing CI/CD test suites?

Developers run their test commands wrapped via agent-mock run -- pytest tests/. The CLI automatically spins up the mock server in a background thread and redirects OPENAI_BASE_URL and ANTHROPIC_BASE_URL to the local twin.