Ask HN: AI agent platform that runs locally

3•EnthusiastShiv•8mo ago

llms are powerful now, but still feel disconnected.

I want small agents that run locally (some in cloud if needed), talk to each other, read/write to notion + gcal, plan my day, and take voice input so i don’t have to type.

Just want useful automation without the bloat. Is there anything like this already? or do i need to build it?

Comments

karanveer•8mo ago

i wanted something similar, especially as im a big advocate of using time blocking method and i use gcal all the time. on top of notion and onenote.

I planned to make something small around this, but currently due to limited time, i am not able to put my hands in it. If you do find a solution please do let me know as well.

EnthusiastShiv•8mo ago

Sure

RobirdLiu•8mo ago

To get started on building it yourself, you could consider a stack like: Python as the base, Ollama for local LLM serving, Whisper (e.g., via whisper.cpp) for STT, Langchain for the agent orchestration and tool integration, and something like Piper for TTS. For Notion/GCal, their official Python SDKs wrapped as Langchain tools would be the way.

This selection prioritizes a pragmatic balance between capability and a more manageable entry point for an individual project, rather than focusing strictly on the absolute SOTA or most experimental options for each component. It should provide a solid foundation for iterative development, and you can always swap out/upgrade specific parts later as you see fit

EnthusiastShiv•8mo ago

yeah this is a solid stack. kinda surprised no one's bundled this up yet into a plug-and-play AI agent starter kit. feels like the kind of thing that should already exist tbh.

RobirdLiu•8mo ago

Thank you so much for your Upvote.This is the first and only Upvote I've ever received.I thought I had answered a few HN Ask questions thoughtfully and seriously,but I ended up with 5 Downvotes.Thank you again,kind soul.I will keep working hard.I still have several great projects that I want to show to the world.In a few more months,perhaps the project I'm currently working on will meet your expectations.However,at the moment,I'm focusing on more fundamental aspects,such as exploring how to maintain continuous and meaningful cognitive activity in LLMs,as well as their ability to reflect and learn.I haven't been working on areas like tool call,memory,TTS,etc.,where there have already been significant achievements.I'll introduce it to everyone once the project is initially functional.Stay tuned!

The Genus Amanita

We have broken SHA-1 in practice

Ask HN: Was my first management job bad, or is this what management is like?

Ask HN: How to Reduce Time Spent Crimping?

KV Cache Transform Coding for Compact Storage in LLM Inference

A quantitative, multimodal wearable bioelectronic device for stress assessment

Why Big Tech Is Throwing Cash into India in Quest for AI Supremacy

How to shoot yourself in the foot – 2026 edition

Eight More Months of Agents

From Human Thought to Machine Coordination

The new X API pricing must be a joke

Show HN: RMA Dashboard fast SAST results for monorepos (SARIF and triage)

Show HN: Source code graphRAG for Java/Kotlin development based on jQAssistant

Python Only Has One Real Competitor

Tmux to Zellij (and Back)

Ask HN: How are you using specialized agents to accelerate your work?

Passing user_id through 6 services? OTel Baggage fixes this

DavMail Pop/IMAP/SMTP/Caldav/Carddav/LDAP Exchange Gateway

Visual data modelling in the browser (open source)

Show HN: Tharos – CLI to find and autofix security bugs using local LLMs

Oddly Simple GUI Programs

The New Playbook for Leaders [pdf]

Interactive Unboxing of J Dilla's Donuts

OneCourt helps blind and low-vision fans to track Super Bowl live

Rudolf Vrba

Autism Incidence in Girls and Boys May Be Nearly Equal, Study Suggests

Wellness Hotels Discovery Application

NASA delays moon rocket launch by a month after fuel leaks during test

Sebastian Galiani on the Marginal Revolution

Ask HN: Are we at the point where software can improve itself?