One OpenAI-compatible API for 250+ models across 40+ providers
Stream-native AI for real-world time series data
Stage
Not given
Startup
Founded
Not given
2026
Based in
US
GB
Pricing model
Freemium
Not given
Pricing
Free: Access all 200+ models with bring-your-own provider keys, or pay-as-you-go on credits. Credits: 5% platform fee on top-ups. Enterprise: Volume discounts, custom routing, full request and response retention (30 days), and 99.9% uptime SLA. Data retention: $0.01 per 1M tokens for Retain All Data
Not given
Key features
Route requests to 250+ models across 40+ providers
OpenAI-compatible API endpoint
Real-time cost and latency tracking
Automatic failover between providers
Request caching
Prompt injection and PII detection guardrails
Self-hosted or cloud deployment
SSO and SAML integration
Spend caps and budget controls
Audit logs and compliance tracking
Anomaly detection and correlation discovery across multiple data streams
Real-time data stream monitoring to flag potential problems
Forecasting by understanding interactions between data streams
Natural language querying with context and assumptions
ci-proto frontier model with improved forecasting accuracy
What makes it different
Full self-hosting under AGPLv3 open-source license
Bring-your-own-keys free forever
Flat 5% platform fee with no minimums or seat charges
Unified analytics across all providers and models
99.9999% effective uptime with automatic provider failover
Stream-native architecture designed specifically for time series data
Foundation models that eliminate context window limitations