Changelog

What we’re shipping.

SDK releases, API updates, and new model additions — in reverse-chronological order.

August 12, 2026 portal/v0.8.0
Starter templates on the developer portal
  • Added /docs/templates with 5 runnable starter projects — chatbot, text summarizer, code reviewer, semantic search, and streaming chat — each with copyable Python and Node.js code.
June 27, 2026 landing/v0.7.0
Enterprise Use Cases section
  • Added PIPEDA Compliance Pack, Financial Services Audit Readiness, and Government Procurement Ready cards to the landing page.
  • Wired docs and quickstart CTAs into the new section so enterprise visitors land in the right place.
June 20, 2026 portal/v0.6.0
Dedicated API key management
  • New /dashboard/api-keys page with full-key reveal (5-minute signed-token expiry) and one-click copy.
  • Regenerate-with-confirmation flow that invalidates the old key the moment a new one is issued.
June 19, 2026 gateway/v0.5.0
Together AI & Fireworks AI routed providers
  • 100+ open models now available: Llama 4, Qwen 3, DeepSeek-V3, GLM-5.2, Mistral, Gemma, Phi.
  • Model-name-based routing — Qwen/GLM/DeepSeek prefer Together, fireworks/* prefers Fireworks, groq/* prefers Groq.
  • /v1/models catalog updated with the full open-model lineup.
June 19, 2026 portal/v0.4.0
Signup and password recovery
  • Forgot / reset password flow with SHA-256-hashed reset tokens and a 1-hour expiry.
  • Signup page creates the account and reveals the raw API key once via a short-lived signed token.
  • /onboard-success page shows the key with copy and a ready-to-run cURL quickstart.
June 18, 2026 gateway/v0.3.0
PIPEDA audit & per-user usage tracking
  • PIPEDA-aligned request audit log: SHA-256 hashed client IPs, raw prompts never stored, 2-year retention.
  • Per-user usage tracking via request_logs; /usage returns aggregates grouped by model.
June 15, 2026 sdk-node/v0.2.0 sdk-python/v0.2.0
Node & Python SDK v0.2.0
  • Streaming responses via Server-Sent Events with backpressure-aware consumption helpers.
  • Built-in retry-with-backoff for transient upstream failures.
  • Supports ?route= strategies (cheapest, fastest, balanced, best_quality) across both SDKs.
June 10, 2026 pricing-page/v0.1.0
Public pricing page
  • Monthly / annual billing toggle with a “Save 20%” badge on the annual option.
  • Tier presets (Open Models, Fast Models, Premium Models) drive an inline input / output token calculator and live cost estimate.
  • Side-by-side comparison table vs. each direct provider.

Build with NorthGate

Read the docs, get an API key, and ship something today.

Read the docs