Damarrion M-H.

probably building something @diizzy_fps

[Damarrion M-H. @__diizzy]

Hey everyone,

For the past few months, I’ve been working on DAEX (Daedalus Execution Engine), an edge-optimized execution client built natively in Kotlin to run local LLMs and autonomous agents directly on Android hardware.

I just pushed it into closed alpha and I’m looking for a few builders to help me stress-test the pipeline, break the local state machine, and see how it holds up across different mobile chipsets.

The app currently features:
- @GoogleGemma models running on #LiteRT (NPU compatibility on select devices)
- Local Hybrid Rag for document search and cross conversational recall
- Local live voice feature (Similar to Gemini Live, but local!)

If you have an Android device (Android 8.0+, ideally with an NPU) and want to test the limits of local on-device agent inference, I'd love your feedback on token generation speeds and pipeline stability.

Repo: https://t.co/zwDWfHpoYw

Join the Closed Alpha here: https://t.co/bnFL7tjR3j

#Gemma4 #Gemma #Android #localAI

[Claude @claudeai]

Introducing Claude Opus 5.

It's a thoughtful and proactive model that comes close to the frontier intelligence of Fable 5 at half the price.

[Damarrion M-H. @__diizzy]

CLAUDEEEEEEE

[Damarrion M-H. @__diizzy]

@xiao18kuma lol, I agree. I kept the pool large because im curious as to what happens if you run this on a device that old.

[Damarrion M-H. @__diizzy]

Introducing DAEX (Daedalus Execution Engine) — a native Kotlin Android app that runs local LLMs + autonomous agents fully on-device. No cloud. No web wrappers.
Pure offline privacy and speed.
Built with:
• LiteRT + Gemma models (Vulkan GPU / NPU delegation)
• Local hybrid RAG + memory
• Offline voice pipeline
• Sandboxed tools + visible reasoning

Just shipped closed alpha — looking for Android builders to stress-test on real hardware (Android 8.0+, NPU preferred) and give feedback on token speeds/stability.

New Discord for testing + on-device AI chats (LiteRT tips, agents, edge perf, etc.): https://t.co/sC8iAFqqxE

Repo: https://t.co/zwDWfHpoYw

Google Group alpha access: https://t.co/bnFL7tjR3j

Who wants to run local agents in their pocket? Drop your device/chipset 👇

#OnDeviceAI #LiteRT #Gemma #AndroidDev #LocalLLM @googlegemma

[Damarrion M-H. @__diizzy]

Introducing DAEX (Daedalus Execution Engine) — a native Kotlin Android app that runs local LLMs + autonomous agents fully on-device. No cloud. No web wrappers.
Pure offline privacy and speed.
Built with:
• LiteRT + Gemma models (Vulkan GPU / NPU delegation)
• Local hybrid RAG + memory
• Offline voice pipeline
• Sandboxed tools + visible reasoning

Just shipped closed alpha — looking for Android builders to stress-test on real hardware (Android 8.0+, NPU preferred) and give feedback on token speeds/stability.

New Discord for testing + on-device AI chats (LiteRT tips, agents, edge perf, etc.): https://t.co/sC8iAFqqxE

Repo: https://t.co/zwDWfHpoYw

Google Group alpha access: https://t.co/bnFL7tjR3j

Who wants to run local agents in their pocket? Drop your device/chipset 👇

#OnDeviceAI #LiteRT #Gemma #AndroidDev #LocalLLM @googlegemma

[Damarrion M-H. @__diizzy]

@Cartidise they need to make a cameraless version/option. It would be so nice to have flagship glasses that are just AR

[Damarrion M-H. @__diizzy]

Repo (open source): https://t.co/zwDWfHpoYw Join the closed alpha (Google Group first → Play Store whitelist): https://t.co/bnFL7tjR3j Who’s in? Reply with your device/chipset if you want early access. First testers will help shape the roadmap.#LocalAI #OnDeviceAI #AndroidDev #Gemma #LiteRT
cc @GoogleDevs

[Damarrion M-H. @__diizzy]

Just shipped the closed alpha of DAEX (Daedalus Execution Engine) — a native Kotlin execution client built from the ground up to run local LLMs + autonomous agents directly on Android hardware.
No web wrappers. No cloud. Full offline autonomy with visible reasoning.

cc
@googlegemma

[Damarrion M-H. @__diizzy]

I’m looking for a small group of Android builders to stress-test the pipeline, break the local state machine, and give feedback on token speeds + stability across different chipsets (Android 8.0+, ideally with NPU).

[Damarrion M-H. @__diizzy]

Hey everyone,

For the past few months, I’ve been working on DAEX (Daedalus Execution Engine), an edge-optimized execution client built natively in Kotlin to run local LLMs and autonomous agents directly on Android hardware.

I just pushed it into closed alpha and I’m looking for a few builders to help me stress-test the pipeline, break the local state machine, and see how it holds up across different mobile chipsets.

The app currently features:
- @GoogleGemma models running on #LiteRT (NPU compatibility on select devices)
- Local Hybrid Rag for document search and cross conversational recall
- Local live voice feature (Similar to Gemini Live, but local!)

If you have an Android device (Android 8.0+, ideally with an NPU) and want to test the limits of local on-device agent inference, I'd love your feedback on token generation speeds and pipeline stability.

Repo: https://t.co/zwDWfHpoYw

Join the Closed Alpha here: https://t.co/bnFL7tjR3j

#Gemma4 #Gemma #Android #localAI

[Damarrion M-H. @__diizzy]

@isidoroflo

[Damarrion M-H. @__diizzy]

@thorwebdev

[Damarrion M-H. @__diizzy]

Hey everyone,

For the past few months, I’ve been working on DAEX (Daedalus Execution Engine), an edge-optimized execution client built natively in Kotlin to run local LLMs and autonomous agents directly on Android hardware.

I just pushed it into closed alpha and I’m looking for a few builders to help me stress-test the pipeline, break the local state machine, and see how it holds up across different mobile chipsets.

The app currently features:
- @GoogleGemma models running on #LiteRT (NPU compatibility on select devices)
- Local Hybrid Rag for document search and cross conversational recall
- Local live voice feature (Similar to Gemini Live, but local!)

If you have an Android device (Android 8.0+, ideally with an NPU) and want to test the limits of local on-device agent inference, I'd love your feedback on token generation speeds and pipeline stability.

Repo: https://t.co/zwDWfHpoYw

Join the Closed Alpha here: https://t.co/bnFL7tjR3j

#Gemma4 #Gemma #Android #localAI

[Damarrion M-H. @__diizzy]

guys gen q, why are we benchmarking every model on svg output. are svgs the next step to agi or am i missing something 🤔

[Damarrion M-H. @__diizzy]

@thorwebdev @HeyGen @GoogleDeepMind @Rames_Jusso @HyperFrames_ Awesome event, the demos were pretty good too! I got some pretty good ideas now👀

[Damarrion M-H. @__diizzy]

@thorwebdev @HyperFrames_ it would appear my presentations are due for an upgrade 👀

[Damarrion M-H. @__diizzy]

@sskarz1016 it hurts my head so much