Prompt Compass

We cut LLM API costs up to 96% with a 0.57 MB router that decides local-vs-cloud in ~5 ms (and blocks PII/jailbreaks).

Get your free API and start saving

[Prompt Compass @promptcompass]

Prompt Compass is officially live on the Microsoft Store.🚀

Windows users can now install the desktop app to add prompt protection, optional local and cloud response routing, and supported AI tool connections through the Prompt Compass gateway.

Available now on Microsoft store.

[Rishabh @RMad985]

Download from Microsoft store -

https://t.co/gVFrNEeRHp

[Rishabh @RMad985]

Real honest numbers.

[Rishabh @RMad985]

Not only route prompts but also block jailbreak attempts and PII detection

[Rishabh @RMad985]

With Qwen 3.8 27b taking over the debate that if frontier models will be running local on device, I have a solution called Prompt Compass where you can easily connect Qwen 3.8 27b from ollama on device and any frontier model’s keys and easily route prompts to save API costs.

[Rishabh @RMad985]

Prompt Compass is now live on the Microsoft Store.

It can protect AI prompts, route supported requests between local and cloud models, and connect with compatible AI tools.

Still a lot more to build, but this is a milestone I wanted to share.

#PromptCompass #AI #MicrosoftStore

[Prompt Compass @promptcompass]

Download Now 👇🏻

https://t.co/XKSoXXsT3y

[Prompt Compass @promptcompass]

Prompt Compass is officially live on the Microsoft Store.🚀

Windows users can now install the desktop app to add prompt protection, optional local and cloud response routing, and supported AI tool connections through the Prompt Compass gateway.

Available now on Microsoft store.

[Prompt Compass @promptcompass]

https://t.co/XKSoXXsT3y

[Rishabh @RMad985]

This is so helpful

[Prompt Compass @promptcompass]

@Kyle_Structure A prompt router that routes the prompt to local edge model or cloud AI, blocks PII and jailbreak attempts all in ~5ms. 0.57mb in size. On device sdk.
Helps vibe coders and businesses running Cloud AI api.

https://t.co/ccE1PaAeTt

[Prompt Compass @promptcompass]

@Joi2James $aving

https://t.co/ccE1PaAeTt

[Rishabh @RMad985]

@nishantsingh211 Wanna save money on cloud AI bills? Check this out 👇🏻

https://t.co/aPwPbiyiTa

[Prompt Compass @promptcompass]

@HeyAliux AI costs:

https://t.co/ccE1PaAeTt

[Prompt Compass @promptcompass]

@riwajrise Routing prompt in under ~5ms to a local or cloud AI model to reduce AI bills (upto 96%), PII detection and Jailbreak —- all in 1 model

https://t.co/ccE1PaAeTt

[Prompt Compass @promptcompass]

@anupamrjp Routing prompt in under ~5ms to a local or cloud AI model to reduce AI bills (upto 96%), PII detection and Jailbreak —- all in 1 model

https://t.co/ccE1PaAeTt

[Prompt Compass @promptcompass]

@anupamrjp Routing prompt in under ~5ms to a local or cloud AI model to reduce AI bills, PII detection and Jailbreak —- all in 1 model

https://t.co/OwmyN383Vu

[Prompt Compass @promptcompass]

@sflorimm Routing prompt in under ~5ms to a local or cloud AI model to reduce AI bills, PII detection and Jailbreak —- all in 1 model

https://t.co/ccE1PaAeTt

[Rishabh @RMad985]

This is just crazy.

[Prompt Compass @promptcompass]

@danellisona https://t.co/WpfRvlE80c

Save upto 96% in cloud AI costs.

[Prompt Compass @promptcompass]

The "safe LLM app" starter pack, 2026 edition:
• embedding router 88 MB
• vector database
• PII scanner
• jailbreak detector, more model
100+ MB and 3 network hops before your prompt reaches any AI.

We fit all four jobs into 0.57 MB. One call. ~5 ms.
https://t.co/ccE1PaAeTt