← FIELD GUIDE

01 / LOCAL AI

Intelligence that
stays with you.

Capture speech. Find your context. Put small models to work locally.

Explore the models ↓
AUDIO→TEXT→MEMORY→ANSWER

CONCEPT / NOT A PRODUCT DEMO

WHY THIS MATTERS

Your context.
Within reach.

Keep useful context close. Local workflows still depend on the complete system.

Read more

Personal AI gets interesting when it can work with the information you actually care about: a thought you spoke aloud, a document you saved, or a decision you want to revisit.

Running parts of that workflow locally creates opportunities for offline use and deliberate control over data movement. The whole experience still depends on model quality, memory, storage, software and the way permissions are designed.

THE MODEL STACK

A useful system
does several jobs.

Different jobs. The right tool for each.

01 / SPEECH

Turn sound
into text.

Whisper turns speech into text on local hardware.

Read more

Whisper is a speech recognition model. whisper.cpp brings Whisper inference to a C/C++ runtime with CPU support, making local transcription a practical area to explore.

Accuracy and speed vary with the model, language, noise and target hardware.

whisper.cpp project ↗
02 / ROUTING

A small decision.
A clear next step.

Needle3 helps route tools and match information.

Read more

Cactus Compute’s Needle3 model card describes function calling and text embeddings. These are useful building blocks for selecting a defined tool or matching information locally.

A model’s suggestion still needs explicit application rules and permissions before an action happens.

Needle3 model card ↗
03 / GENERATION

Summarize.
Draft. Revisit.

Explore local writing and summaries with llama.cpp.

Read more

Small instruction models can be explored for focused writing and summarization tasks. llama.cpp provides local inference across a range of hardware, including CPUs.

Model fit involves more than the parameter count: weights, context and runtime overhead all consume memory.

llama.cpp project ↗

PRIVATE CONTEXT

Memory you
can trace.

Find useful context before asking the model.

Read more

Retrieval lets an application find relevant material before asking a model to answer. It can combine text search with embeddings, which represent information for similarity matching.

01

Keep the source

Keep audio, transcripts and documents as evidence.

Read more

Audio, transcripts and documents remain the reference for what happened.

02

Find the relevant passage

Find relevant passages in your archive.

Read more

Search selects useful context instead of sending an entire archive through a model.

03

Return to the evidence

Revisit the source. Correct misunderstandings.

Read more

An answer should make it easy to revisit its source and correct a misunderstanding.

ONCORE / IN DEVELOPMENT

Personal AI, still in development.

Read more

OnCore’s direction brings private capture, local memory and user-controlled assistance into one personal system. These are areas of development; final capabilities and hardware will be shared as they are validated.

THE TRADE-OFFS

What makes
a model fit?

Memory

Weights and working context must fit together.

Read more

Weights, working context and other services need room together. A file that fits on disk may still exceed the available RAM.

Responsiveness

Measure the complete interaction, from input to answer.

Read more

Measure the complete interaction: loading, transcription, retrieval and generating the answer you need.

Power & heat

Useful work must fit the device’s energy budget.

Read more

A pocket-sized device has a different energy budget from a desktop. Useful work must fit its operating environment.

Runtime support

The model and hardware backend must work together.

Read more

ONNX Runtime uses execution providers for CPUs and supported accelerators. The exact model and backend must work together.

ONNX Runtime documentation ↗

YOUR SIGNAL

What matters
to you?

Choose your future updates.

Choose my updates ↗NEXT / CHIPS & DEVICES →

YOUR BRIEFING

Follow your curiosity.

Choose topics. Sign in to save them.

Your topics

Email delivery is not connected yet.

Preview my briefing ↗

YOUR AI BRIEFING

Make this
your signal.

Create a profile to keep your topic choices.

Forgot your password? ↗

Your email saves your choices. Change topics or turn email off anytime.

Email confirmation and delivery are opening soon.