Multi-Provider AI Gateway for IDEs, Coding Agents & Local AI
AlphaOne HyperRouter is a Windows-native AI gateway that connects compatible IDEs, coding assistants, CLIs, and agent harnesses to multiple cloud AI providers and Local GGUF models through a single local endpoint.
It provides unified OpenAI Chat Completions, OpenAI Responses API, and Ollama-compatible interfaces with automatic provider routing and failover.
Official public distribution repository for AlphaOne HyperRouter binaries and release notes. The HyperRouter source code is proprietary and is not distributed through this repository.
Download the latest official Windows release from:
All published versions are available on the:
Each release includes its own version notes and downloadable binary package.
HyperRouter runs locally on Windows and provides one endpoint between your AI client and multiple AI providers.
IDE / Coding Agent / CLI
│
│ OpenAI / Responses / Ollama Protocol
▼
AlphaOne HyperRouter
│
├── Local GGUF AI
├── OpenAI
├── Anthropic
├── Google Gemini
├── DeepSeek
├── Qwen
├── Kimi AI
├── Agnes AI
└── Groq
Instead of configuring a separate provider connection in every compatible client, HyperRouter provides a stable local gateway and manages provider selection centrally.
Connect multiple supported cloud providers using your own API keys and manage them from one local HyperRouter installation.
Supported provider integrations include:
- OpenAI
- Anthropic
- Google Gemini
- DeepSeek
- Qwen
- Kimi AI
- Agnes AI
- Groq
When an eligible provider or model encounters supported quota, rate-limit, or availability failures, HyperRouter can automatically continue routing through the next configured provider or cascade tier.
Run compatible .gguf models locally without requiring a cloud API.
HyperRouter supports local execution using available:
- NVIDIA CUDA acceleration
- Vulkan acceleration
- CPU fallback
Local AI remains optional and is managed separately from cloud provider routing.
Default OpenAI-compatible endpoint:
http://127.0.0.1:11436/v1
Primary Chat Completions endpoint:
POST /v1/chat/completions
OpenAI Responses API endpoint:
POST /v1/responses
For compatible clients using the Ollama protocol:
http://127.0.0.1:11436
Supported compatibility routes include:
/api/chat
/api/generate
/api/tags
HyperRouter maintains canonical conversation and tool-call history so compatible agent workflows can preserve task context across supported provider transitions.
Provider-native state is used where appropriate while canonical history remains the portability layer across providers.
Optional conservative context optimization reduces unnecessary formatting overhead for supported cloud requests while prioritizing content integrity.
The feature can be disabled at any time if unusual tool behavior occurs.
HyperRouter provides local telemetry for:
- Input tokens
- Cached input tokens
- Output tokens
- Reasoning / thought tokens
- Tool-use tokens
- Total tokens
- Request latency
- Time to first token
- Provider and model
- HTTP status
- RPM / TPM / RPD quota classification
Provider API keys are managed locally by HyperRouter.
On supported Windows installations, provider credentials are protected at rest using Windows DPAPI under the current Windows user profile.
Existing provider credentials are not rendered back into the Settings page.
HyperRouter is designed for clients that support OpenAI-compatible, Responses API, or Ollama-compatible endpoints.
Validated integration patterns include:
- Visual Studio Code
- Visual Studio
- OpenAI Codex CLI
- OpenAI Codex IDE Extension
- Cline
- Roo Code
- Cursor
- Continue.dev
- Android Studio
- Compatible agent SDKs and harnesses
The IDE or coding client remains responsible for filesystem access, terminal execution, editing, approvals, and other local agent actions.
HyperRouter provides the AI routing, protocol translation, provider access, and continuity backend.
Go to:
Download the Windows package and extract it if required.
Run:
AlphaOne.AI.Bridge.exe
The default local server address is:
http://127.0.0.1:11436
Open:
http://127.0.0.1:11436/setting
Configure at least one cloud provider API key or enable a compatible Local GGUF model.
For OpenAI-compatible clients:
Base URL : http://127.0.0.1:11436/v1
API Key : sk-alphaone
Model ID : alphaone-hyperrouter
The API key shown above is only a placeholder for clients that require a non-empty credential field.
Actual upstream provider credentials remain configured inside HyperRouter.
For Ollama-compatible clients:
Base URL : http://127.0.0.1:11436
Model : alphaone-hyperrouter
Compatible clients can use one stable public model ID:
alphaone-hyperrouter
The actual provider model class is selected centrally from the HyperRouter Settings dashboard.
Available profiles:
| Profile | Purpose |
|---|---|
| High | Maximum-capability provider model class |
| Medium | Balanced capability and efficiency |
| Low | Faster / efficiency-oriented model class |
| Custom | Uses individually configured provider model selections |
Provider priority and model profile are independent.
The selected provider order determines where HyperRouter routes the request, while the model profile determines which model class is selected for that provider.
Every public release is published manually through GitHub Releases.
Release pages contain:
- Version number
- Release date
- Change summary
- Bug fixes
- Compatibility changes
- New features
- Downloadable Windows binary/package
See:
Older versions remain available for reference and rollback where applicable.
HyperRouter is designed primarily as a local single-user Windows gateway.
Default configuration:
Address : 127.0.0.1
Port : 11436
HyperRouter includes local browser-origin restrictions and network binding controls intended to prevent unintended remote access under the default configuration.
Users remain responsible for:
- Protecting their provider API credentials
- Downloading HyperRouter only from official AlphaOne AI distribution channels
- Keeping Windows and provider credentials secure
- Reviewing release notes before upgrading production environments
- Windows
- Network access for cloud AI providers
- Provider API key(s) for cloud AI usage
- Optional compatible
.ggufmodel for Local AI - Compatible IDE, coding assistant, CLI, or API client
Actual Local AI performance depends on model size, available RAM, GPU VRAM, and hardware acceleration support.
AlphaOne AI
HyperRouter product information:
https://alphaone-ai.com/hyperrouter.html
For HyperRouter questions, beta testing, compatibility reports, or release-related issues:
When reporting an issue, please include where applicable:
- HyperRouter version
- Client / IDE name
- Protocol used
- Provider involved
- HTTP status code
- Relevant HyperRouter log information
Do not include API keys or other credentials in public GitHub issues.
This repository is intended for:
- Official AlphaOne HyperRouter release announcements
- Version history
- Release notes
- Documentation
- Public binary downloads
It is not the HyperRouter source-code repository.
Executable files published under the official Releases section are the intended public distribution artifacts.
AlphaOne HyperRouter is proprietary software.
Copyright © 2026 AlphaOne AI. All rights reserved.
The availability of downloadable binaries through this repository does not grant permission to reproduce, redistribute, reverse engineer, modify, sublicense, or commercially exploit the software except where expressly authorized by AlphaOne AI.
Third-party product names, trademarks, APIs, and services referenced by HyperRouter remain the property of their respective owners.
AlphaOne AI is an independent third-party software provider and is not affiliated with, endorsed by, or sponsored by OpenAI, Anthropic, Alibaba Cloud, Moonshot AI, Google, Agnes AI, Groq, or other referenced AI providers.
One Router. Multiple Providers. Local + Cloud AI.