Host local AI models for your apps from one shared model library. Start on this PC with Free, serve other devices with Pro Personal, or add API keys and governance with Pro Commercial. AI Server turns this PC into a shared AI runtime for every AI Suite app: one model download and one model library, used everywhere. The Free tier is permanent, not a time-limited trial. It provides unlimited local AI for AI Suite apps on this PC, but is limited to loopback hosting; generic non-AI-Suite clients are capped at 1 request per minute and 10 requests per UTC day, and Free does not include network serving, service mode, custom server settings, usage analytics, API keys, or governance. Pro Personal unlocks network serving, Windows service mode, server settings, usage analytics, and unmetered generic-client access for personal use. Pro Commercial additionally unlocks API keys, quotas and rate limits, content filtering, the multi-host gateway, signed audit export, model-lifecycle controls, regional-compliance settings, and organization enrollment for business use. Telemetry is content-free; server access keys are stored only as hashes; provider credentials are encrypted locally. Requests stay on your device when you use local models, while optional cloud models use the provider endpoint you configure. Free: Run AI Server as a loopback-only local server for AI Suite apps on this PC — manage and download models, connect providers, and keep older AI Suite clients working. AI Suite app access is unlimited; generic clients are capped at 1 request per minute and 10 requests per UTC day. The Free tier does not expire. Pro Personal (subscription — in-app purchase): Serve other devices on your network, run AI Server as a background Windows service, choose the port and display name, view usage analytics, and remove generic-client caps. API keys and governance require Pro Commercial. Pro Commercial (subscription — in-app purchase): Adds governance — per-key rate limits, usage quotas, content moderation, the multi-host gateway, signed audit export, model-lifecycle policy — plus organization enrollment and unlimited API keys. AI Server supports fully local processing with downloaded models, including hardware-accelerated models on supported Windows devices, and no internet connection. Its live activity page shows current and recent requests—including the effective client IP for operator diagnosis—without displaying prompts or responses. If you choose to configure a cloud provider, requests using that provider are sent to its configured endpoint.
Productivity · Windows · Free
