Changelog

RSS Feed

Everything we've shipped — models, API features, platform improvements, and security hardening. Subscribe to the blog for major announcements.

Apr 2026GA Releasev1.0
  • GPU cost optimization engine: spot instance scheduling, preemption handling, automatic fallback to on-demand
  • Autoscaling policies: time-based, load-based, and budget-aware scaling rules
  • Batch pricing tier: up to 40% discount on asynchronous workloads
  • Usage alerts: configurable thresholds (50/75/90/100%) with email and webhook delivery
  • Spending limits dashboard: monthly budget caps, real-time burn-rate progress bar
  • Alert history log with status tracking (sent, acknowledged, resolved)
Apr 2026Developer Experiencev0.9.5
  • Webhook system: subscribe to inference.completed, usage.threshold, and billing.invoice events
  • Webhook signatures: HMAC-SHA256 on every payload — verify with the X-Cloudach-Signature header
  • Onboarding email sequence: 7-touch drip campaign triggered on signup
  • Model selection guide and use-case matrix published to /docs
  • Interactive API playground embedded in /docs — run live requests without leaving the page
  • Improved dashboard error states with actionable recovery suggestions
Mar 2026Securityv0.9
  • Auth hardening: bcrypt password hashing, JWT-based sessions, secure cookie flags
  • Input validation and SQL-injection prevention across all API routes
  • Rate limiting (60 RPM / 1M TPD) with sliding window and Retry-After header
  • CORS policy tightened; X-Content-Type-Options, X-Frame-Options headers added
  • API key stored as SHA-256 hash — raw key never persisted after creation
Mar 2026Enterprisev0.8.5
  • Enterprise page redesigned: SLA table, compliance badges, dedicated support callout
  • Team management dashboard: invite members, assign roles (admin/developer/viewer), revoke access
  • Fine-tuning workflow: LoRA adapter training, dataset upload, job queue, deployed adapter endpoints
  • E2E test suite (Playwright): 18 critical-path tests covering auth, API keys, models, and billing
  • SDK packages: official cloudach-python and cloudach-node on PyPI and npm
Mar 2026Infrastructurev0.8
  • Autoscaling from zero: GPU pods scale to 0 when idle, cold-start in ~30s
  • Tenant isolation: per-tenant Kubernetes namespaces and network policies
  • CI/CD pipeline: GitHub Actions → Docker build → GKE deploy on every merge to main
  • Observability: Prometheus metrics, Grafana dashboards, PagerDuty alerting
  • Multi-region deployment: US-east, EU-west routing with CDN edge caching
Feb 2026Performancev0.7.5
  • Dashboard performance: self-hosted fonts, HTTP caching headers, lazy-loaded charts
  • CLI v0.1: cloudach chat, cloudach models list, cloudach keys create/revoke
  • Monitoring and alerting: structured JSON logs, Datadog integration, custom dashboards
  • API error reference: all error codes documented with causes and remediation steps
  • Rate limit docs expanded: per-tier limits, burst budgets, and backoff examples
Feb 2026Dashboardv0.7
  • API Keys page: create, name, copy, and revoke keys with one click
  • Usage page: daily token consumption charts, per-model breakdown
  • Models page: model catalog with context window, pricing, and status indicators
  • Settings page: account management, password change, danger zone
  • ErrorBoundary: graceful fallback UI for dashboard render errors
Feb 2026APIv0.6
  • POST /v1/chat/completions: OpenAI-compatible chat endpoint with streaming (SSE)
  • POST /v1/completions: legacy text completion endpoint
  • GET /v1/models and GET /v1/models/:id: model catalog endpoints
  • GET /health: unauthenticated health check endpoint
  • Streaming: Server-Sent Events with delta chunks and [DONE] terminator
Feb 2026Modelsv0.5
  • Llama 3 8B deployed on vLLM — 8K context, sub-100ms median TTFT
  • Llama 3 70B available on dedicated GPU tier
  • Mistral 7B: 32K context window, EU-hosted option
  • Mixtral 8×7B: mixture-of-experts, highest accuracy tier
  • Benchmark results: 96ms p50 TTFT on Llama 3 8B at 10 RPS
Jan 2026Platformv0.4
  • Neon Postgres integration: connection pooling, SSL-only, combined schema migration
  • Auth system: /api/auth/register, /login, /logout with secure session management
  • API gateway service: request routing, auth middleware, token metering
  • Vercel deployment: serverless Next.js frontend, edge config for API base URL
Jan 2026Launchv0.1
  • Initial commit: Next.js frontend, Kubernetes manifests, Docker configuration
  • Brand identity: Cloudach logo, indigo/white design system
  • Public landing page with pricing, feature highlights, and CTA
  • Blog: first post on the product vision and developer-first approach
API DocsBlogStatusRSSsupport@cloudach.com