One API for the AI frontier

One API.
Multiple models.

AI Ranger brings independently deployed AI projects into one catalog. Start with ModelRouter to browse supported models and route multi-provider inference through one connection.

Get a ModelRouter API key Explore AI Ranger projects

Browse the ModelRouter catalog before signing in. Chat, probes, and API-key management require an account.

Ranger stage

5 signals worth following

One app. Your full attention. Scroll to explore, or choose a signal below.

0105

LivePrimary application

AiRanger ModelRouter

One connection. Many models. Intelligent routing.

The primary AI Ranger workspace: manage API tokens, discover and probe models, bring your own provider keys, and run chat or streaming inference through one connection.

Explore app

Current ModelRouter console overview in light mode, with an API quickstart, workspace setup steps, and real Llama, DeepSeek, and Qwen model cards available to browse before sign-in. — Public product surface · 2026-10-14The preview illustrates one API connection routing to model providers, followed by managing API tokens, discovering and probing models, and chat or streaming inference. Sign-in is required in the production workspace; no requests are sent here.

0205

LiveField report 02

Dendrite

One API. Many machines.

A distributed inference control plane that routes OpenAI-compatible requests across Linux, CUDA, Windows, WSL2, CPU, and GPU providers.

Explore app

Dendrite interface showing a network topology connected to distributed inference providers. — Public product surface · 2026-09-21The preview illustrates the Dendrite router connected to CUDA, Linux, Windows, CPU and WSL2 providers. Provider Forge guides hardware discovery, runtime and model preparation, operator-invite enrollment, and local serving.

0305

AlphaField report 03

Elastic Inference Fabric

Keep inference where the hardware lives.

An authenticated control surface for observing local capacity, routing work deliberately, and connecting agents to private inference nodes.

Explore app

Public Elastic Inference Fabric demo showing an agent routed through Cloudflare to a local inference node. — Public product surface · 2026-09-21The architecture preview shows an agent sending a task through the authenticated Cloudflare MCP router to a local model, with heartbeats and results returning. An administrator can choose to share results with the Ambiguous workspace. This is not live telemetry.

0405

LiveField report 04

AI Labs San Diego

Give useful hardware a second life.

Offer working compute hardware or e-waste for San Diego learning labs and community AI projects. A person reviews each offer and confirms any drop-off.

Explore app

AI Labs San Diego public hardware donation homepage with current needs, intake steps, and an offer-hardware link. — Public product surface · 2026-10-20The October 20 public screenshot shows hardware donation intake. The following illustrations explain offering hardware, human review, and a confirmed drop-off. No offer is submitted by this preview.

0505

Design previewField report 05

RecruiterAgent

Evidence-led resume review. Human decisions.

Explore candidate evidence, requisition criteria, and review tasks in an OpenClaw skill workspace. This design preview uses synthetic data; no model or OpenClaw connection is active.

Explore preview

RecruiterAgent design preview showing synthetic candidates, evidence criteria, and human review controls. — Public product surface · 2026-10-15The design preview shows synthetic candidate submissions, evidence against requisition criteria, and review tasks for human decisions. No model or OpenClaw connection is active; nothing is sent from this portal preview.

Discovery field

The complete catalog

Choose ModelRouter for supported model APIs, Dendrite for distributed inference, or AI Labs San Diego for community hardware reuse. Elastic Inference Fabric requires authentication; RecruiterAgent is a synthetic design preview. Each application has its own access requirements.