Developer Hub
Bring quickstart, public docs, the live model catalog, interactive playground, pricing, and machine-readable agent resources together in one developer entry.
Drop-in base_url migration with native VaaS on-chain proof & Agent loop protection
# 1. Install official OpenAI SDK: pip install openai
from openai import OpenAI
# 2. Point base_url to BatchIn Inference Cloud
client = OpenAI(
api_key="YOUR_BATCHIN_API_KEY",
base_url="https://api.batchin.tech/v1",
default_headers={
"X-VaaS-Anchor": "true", # Base L2 Merkle proof commitment
"X-Agent-Run-Id": "run_live_demo_01", # Multi-step session tracking
"X-Agent-Max-Budget": "1.50", # Hard spend ceiling ($1.50 cap)
}
)
# 3. Dispatch inference through documented OpenAI-compatible routes
response = client.chat.completions.create(
model="deepseek-v4-pro",
messages=[
{"role": "system", "content": "You are a senior Silicon Valley AI infrastructure architect."},
{"role": "user", "content": "Explain why VaaS cryptographic receipts prevent enterprise AI vendor repudiation."}
],
temperature=0.2,
stream=True,
)
for chunk in response:
content = chunk.choices[0].delta.content
if content:
print(content, end="", flush=True)Documented routes support the compatible OpenAI parameters, tool calls, and streaming behavior listed in the API contract.
Every request is signed with Ed25519 and batched into Base L2 Merkle tree roots.
Gateway-level loop mitigation, hard budget ceiling cuts, and semantic 429 fallbacks.
Global Access, Unified Ingress
BatchIn coordinates low-latency API gateways, security controls, and high-density compute clusters to deliver unified enterprise engineering standards.
Open docsUltra-low latency inference, global edge ingress, USD pricing, and self-serve API keys.
Enterprise SLAs, high-concurrency video generation pipelines, and 24/7 dedicated engineering support.
Edge ingress
Latency work starts at the customer-facing edge before requests enter the primary execution path.
Streaming delivery
Connection reuse and consumer isolation are tuned to reduce jitter and long-tail failures.
Traffic policy
Customers see a simple API and clear limits while BatchIn handles traffic protection behind the scenes.
Public contract and readiness
Shared core
This is where public Model API, batch, usage, billing, and public MCP contract stay consistent.
Private lanes
Customer UI keeps one product truth while delivery, quota, and isolation can vary by contract.
Capacity truth
Public pages show inventory and availability only from the verified capacity registry.
Production posture
Built for stable production-scale traffic.
Traffic mix
Text plus vision, audio, image, and video workloads.
Streaming path
Regional ingress, stable streaming, and request continuity.
Control guardrails
Scoped limits, request isolation, and backpressure controls.
Chat Completions
Use the familiar chat path for low-friction migration, streaming responses, and production request telemetry.
Responses API
Move tool-oriented and structured application flows onto the newer responses surface without splitting billing or traces.
Embeddings & Vectors
Vector and similarity endpoints sharing unified routing, usage, and billing metadata.
Media & Multimodal
Image, speech, transcription, and video endpoints with cryptographic VaaS receipts.
Agent Discovery & MCP
Expose MCP, llms.txt, agents.md, and machine-readable discovery links from one developer entry.
Full OpenAPI specifications and endpoint parameter definitions.
View DocsExplore the live public models with context windows and pricing.
Browse ModelsTest chat completions, structured outputs, and media tasks in real time.
Open PlaygroundSimple pay-as-you-go rates and enterprise volume capacity reservations.
View PricingOfficial client libraries for TypeScript, Python, Go, and OpenAI SDK drop-in.
Explore SDKsPatterns for building autonomous tool-calling swarms and durable workflows.
Learn More