Frontier output, a fraction of the tokens. A yes/no decision doesn't need a frontier model. Neither does checking whether a document is relevant. @smartloop sends each part of a request only as far as it needs to go. 01 Split the request A request becomes small typed steps: collect, judge, extract, tool, fetch, reason. Each step gets only the context it needs. No step gets the whole conversation. 02 Spend tokens where they count Decisions run on the fast System 1 model, capped at 64 tokens. Tool resolution gets room to reason. Only content steps load a bigger capability model. 03 Filter before you prompt Retrieved context is judged for relevance before it enters a prompt. An unrelated document costs zero tokens. 04 Share weights, skip the bill One model host keeps weights loaded across every project. If a model is skipped, the stream says why. Here is what local first means. If your machine has the resources and the model, the step runs on your machine. If it doesn't, the step is delegated to a cloud model. Open and frontier models run as one system. On real hardware: Contract review on an M4 Mac, 24GB. Orchestration, hybrid search, model selection and eval on a five-year-old Linux machine running Qwen. Built in Rust on llama.cpp. Quantized open models on macOS, Linux and Windows. MCP tools, reusable skills, agents, OpenAI-compatible API. Studio 1.0.58 added Legal and Healthcare templates, including contract review, plus one-click MCP connections. It started as private document processing for European financial institutions. Embedding every document into a single model hit its limits, so I split the work into steps. That was two years ago, mostly nights and weekends. I'm looking for a few enterprise design partners in privacy-sensitive industries, starting with finance and law. Bring a real workflow and we'll build around it. #LocalAI smartloop.ai/docs

Frontier quality. Local First @smartloop smartloop.ai/?s=x

If you're using doc-to-chat wrappers and sending your personal and legal documents to a product you don't know, do yourself a favor: download @smartloop from the App Store, pick a #Qwen model from the list, add your documents, set up your MCP and skills, and optionally enable AI search. All of this costs $0 (unless you need AI search <=$5), and your data stays with you. #DocToChat

#AI search with @Smartloop Studio — @Google's Gemma4 (E2B) + @brave , running 100% locally on @Apple silicon. The demo: product recommendations from an image or a plain-text query. 0 tokens, and Brave's Search API is free up to 1,000 requests/month. #LocalAI #aisearch

Dropped support for @MistralAI , now base model use multiple models in a single prompt to product best result based on speed and model capabilities #localAI #aiorchestration, so far very impressed with ministral-3-3b @smartloop P.S. All my tests are done in base m4-mac with 24GB memory @Apple

Two new flavors for the dashboard to show how much tokens being saved #localAI

#LocalAI and foundation model improvements could move FMCG to the next level, where brands and pre-release products can stay in-house while boosting the R&D phase. Here is a demo on how to identify nail color and thus recommend or test products.

Smartloop Studio is now available for #windows, it's your local on-deivce co-pilot for reasoning and extracting information. Instead of using one single model it uses multiple models to drive a solution which all run locally and within a given memory constraint to produce comparable results. The GPU acceleration is supported on Vulkan compatible devices (#intel, #AMD or #nvidia) @Microsoft Download it here: smartloop.ai/download Here is a quick preview on windows on how I am extracting detailed information from vanguard fund prospectus:

Two flavors of official theme @smartloop

IF you are looking for a gpt mac desktop app that runs locally and produces comparable result , look no further download @smartloop #AI #local #agent. We use #gemma @Google family of models to fine-tune and orchestrate smartloop.ai Its free too

Local AI agents at work #localAI #agent #orchestration youtu.be/yVLH7uJrbTk

Here is a quick overview of #AI agent harness using small models on #macos @apple #LocalLLaMA Steps: 1. Open terminal (macos) 2. Follow the instruction as outlined in the readme: github.com/smartloop-ai/smar… 3. Type slp (to start the harness) For desktop app, here is the link to beta requests: smartloop.ai/ Here is quick uncut clip using @screenstudio

Here at a @MacPaw event , just heard the keyword “Apple hardware is great for running local models.” Can’t agree more , easy to install and less dependencies @Apple

Experimenting with github style contribution chart for token / second metrics #LocalLLaMA