Smartflow 1.7

AI Packaging Platform

Ship your own branded AI API in a day · Real-time analytics across every request · Provider Onboarding Wizard · Hybrid Data Bridge · Model-by-model performance

AI Packaging Provider Wizard Analytics Overhaul Hybrid Bridge Live Metrics
Release Summary
#FeatureCategoryImpact
1AI Packaging Platform — new operating mode for white-label AI APIsNew FeatureNew GTM motion · ISV, MSP, enterprise
2Provider Onboarding Wizard — 5-step guided UI in Routing dashboardNew FeatureZero-code provider setup in <5 min
3Hybrid Analytics Bridge — new service merging Redis + MongoDB for complete analyticsInfrastructureUnifies 119K+ request history
Platform & Product
1
AI Packaging Platform ★ Headline Feature
New Feature

Smartflow 1.7 introduces a new operating mode: the AI Packaging Platform. Any organisation can now take one or more LLM providers — OpenAI, Anthropic, Google Gemini, Ollama, or any OpenAI-compatible endpoint — and re-expose them as a fully governed, branded AI API under their own domain and their own rules.

Your customers call https://ai.acme.com/v1/chat/completions, use model names like acme-premium, and authenticate with sk-acme-xxxx virtual keys — while Smartflow handles everything behind the scenes: routing, caching, compliance, identity, metering, and the audit log. They never know which underlying provider is serving their request.

This enables three entirely new go-to-market motions:

  • ISV / SaaS: Ship your AI product layer in a day instead of months, with built-in caching that cuts provider costs by 30–60%.
  • MSP / AI Reseller: Provision per-client AI environments with isolated virtual keys, usage quotas, and compliance policies — all managed centrally.
  • Enterprise Platform Team: Expose a governed internal AI API to business units and partners, with an immutable per-request audit trail and SSO identity integration.
Enables

White-label AI APIs · Per-tenant isolation · Provider swap without client code changes · Full compliance + audit at the gateway layer

→ Full use case guide: AI Packaging Platform documentation

2
Provider Onboarding Wizard
New Feature

The new Provider Onboarding Wizard — available in the Routing dashboard under the green "Onboard Provider" tab — guides operators through provider setup in 5 steps with zero YAML or config file editing.

  • Step 1 — Choose Provider: 8 pre-configured tiles: OpenAI, Anthropic, Google Gemini, Azure OpenAI, Ollama, AWS Bedrock, Cohere, and Custom (any OpenAI-compatible endpoint). Auto-fills base URL and auth method.
  • Step 2 — Configure: Masked API key field with show/hide toggle, editable base URL, org ID (OpenAI), Azure deployment name, or AWS region — only shows fields relevant to the selected provider.
  • Step 3 — Test Connection: Validates credentials and endpoint reachability. Shows discovered models, latency, and a clear pass/fail indicator.
  • Step 4 — Routing Policy: Set as default, fallback, or temporary override (with configurable TTL). Strategy weight sliders (success rate, latency, cost, quality). Model name mapping builder.
  • Step 5 — Review & Activate: Full configuration summary plus one-click activation that calls PUT /api/routing/strategy, PUT /api/routing/default-provider, and POST /api/routing/model-mapping via the Smartflow API.
Enables

Any operator can onboard a new AI provider in under 5 minutes — no API calls, no config files, no YAML required.

Analytics & Observability
3
Hybrid Analytics Bridge — Complete Dashboard Visibility
Infrastructure

Smartflow 1.7 ships the Hybrid Data Bridge — a new service that gives the analytics dashboard complete visibility across your full request history, not just the live window. Cache performance, model-by-model routing metrics, request trends, and provider breakdowns now populate automatically from day one of deployment with no manual configuration.

The bridge delivers:

  • Real-time cache metrics — cache hit rate, tokens saved, and cost savings visible in the Cache Stats dashboard
  • Model-by-model routing performance — per-model latency, success rate, and request volume across all configured providers
  • 30-day request trend — historical daily traffic chart for capacity planning
  • Full provider breakdown — per-provider token usage, cost, and success rate in the analytics pages
Enables

Complete analytics visibility from day one · No data migration · Works with existing Redis and MongoDB deployments

Upgrading to 1.7
↑
Upgrade Notes

Kubernetes deployments: Apply the Hybrid Bridge deployment manifest to add the new service:

kubectl apply -f https://docs.aperion.ai/assets/hybrid-bridge-deploy.yaml

Then update the main ingress to route /api/vas/metrics to the bridge service (see deployment guide). All other services continue with a standard image update:

docker pull langsmartai/safechat-enterprise:v1.7.15b
docker pull langsmartai/safechat:v1.7.1
docker pull langsmartai/safechat-dashboard:v1.7.1
docker compose -f docker-compose.simple.yaml up -d --pull always

No database migrations required. The Hybrid Bridge reads the existing MongoDB smartflow_vas collection and Redis keyspace without schema changes. Provider Onboarding Wizard is included in the dashboard image.

New Kubernetes resource: smartflow-hybrid-bridge Deployment + Service + Ingress rules for /api/redis, /api/hybrid, and /api/vas/metrics.