Live Requests

GET /v1/models/list

8s ago

Served from cache (0.5ms)

POST /v1/embeddings

14s ago

Routed to iad02-edge in 19ms

POST /v1/chat/completions

31s ago

Failover triggered → sin01-edge

req_9f8a7b6c…

POST /v1/chat/completions

Target: gpt-4o

LATENCY

12ms

EDGE NODE

fra01

STATUS

200 OK

{“status”: “routed”, “latency_ms”: 12, “model”: “gpt-4o”}

Live Requests

GET /v1/models/list

8s ago

Served from cache (0.5ms)

POST /v1/embeddings

14s ago

Routed to iad02-edge in 19ms

POST /v1/chat/completions

31s ago

Failover triggered → sin01-edge

req_9f8a7b6c…

POST /v1/chat/completions

Target: gpt-4o

LATENCY

12ms

EDGE NODE

fra01

STATUS

200 OK

{“status”: “routed”, “latency_ms”: 12, “model”: “gpt-4o”}

Live Requests

GET /v1/models/list

8s ago

Served from cache (0.5ms)

POST /v1/embeddings

14s ago

Routed to iad02-edge in 19ms

POST /v1/chat/completions

31s ago

Failover triggered → sin01-edge

req_9f8a7b6c…

POST /v1/chat/completions

Target: gpt-4o

LATENCY

12ms

EDGE NODE

fra01

STATUS

200 OK

{“status”: “routed”, “latency_ms”: 12, “model”: “gpt-4o”}

Engineered for the Enterprise.

Mission-critical infrastructure routing billions of tokens with zero-latency failover.

4.2 Billion Requests

Routed Monthly

8ms P99 Latency

Meridian

8ms P99 Latency

Meridian

8ms P99 Latency

Meridian

1.2B Tokens / mo

Forrge Labs

1.2B Tokens / mo

Forrge Labs

99.997% Uptime, Trailing 12mo

Lucid AI

99.997% Uptime, Trailing 12mo

Lucid AI

0 Unplanned Failovers, Year to Date

Ledger Peak

0 Unplanned Failovers, Year to Date

Ledger Peak

14 Edge Regions, Zero Cold Starts

Flowstack

14 Edge Regions, Zero Cold Starts

Flowstack

14 Edge Regions, Zero Cold Starts

Flowstack

DEVELOPER EXPERIENCE

Integrate in minutes. Scale to infinity.

Drop-in SDKs for Python, Node, and Go. Type-safe responses and native edge-caching out of the box.

# Synapse OS Python inference example
from synapse import SynapseClient

client = SynapseClient(api_key="sk_live__xxx")
response = client.inference.create(
    model="reasoner-v2",
    input="Summarize this deployment log."
)
print(response.output_text)
View on GitHub

CORE INFRASTRUCTURE

Built on bare metal. Deployed at the edge.

Stop worrying about regions. Our anycast network automatically routes your payloads to the closest active GPU cluster in under 15ms.

≤ 15 ms

Client

Edge PoP

GPU cluster

Anycast edge routing

Every request lands on the nearest healthy GPU cluster automatically — no region config, no cold starts.

≤ 15 ms

Client

Edge PoP

GPU cluster

Anycast edge routing

Every request lands on the nearest healthy GPU cluster automatically — no region config, no cold starts.

Enterprise-grade security

Synapse OS is SOC-2 Type II and HIPAA compliant. Bank-grade encryption applied automatically to every edge payload.

Enterprise-grade security

Synapse OS is SOC-2 Type II and HIPAA compliant. Bank-grade encryption applied automatically to every edge payload.

Real-time telemetry

Every spike absorbed, every region visible. Export to Datadog, Axiom, or S3.

1,204 req/s
edge-tyo
edge-fra
edge-nyc
edge-sgp

Developer-first infrastructure.

DEVELOPER ECOSYSTEM

Comprehensive API references, native SDK guides, and transparent release notes. Everything your engineering team needs to integrate in minutes.

COMMAND MENU

Search across docs, APIs & guides

Quickly locate endpoint references, SDK docs, and security configurations across the Synapse network.

Search documentation…

⌘K

Live · synced from CMS

DEPLOY / SYNAPSE OS

Start building with Synapse OS.

Two ways in — pick your path.

01

For humans

Get started with our Next.js and Python SDKs in seconds.

02

For agents

Tools to connect your autonomous agents to our edge infrastructure.

$npm install @synapse/sdk

Create a free website with Framer, the website builder loved by startups, designers and agencies.