Every time you start a new conversation with an AI assistant, you usually have to re-explain what you're working on — paste in the doc, describe the meeting, remind it what project this is about. minimi's idea is to remove that re-explaining entirely: it quietly watches what's on your screen and stores it as memory, so Claude already has the context of what you've been reading, discussing, and working on. minimi is a macOS app that passively captures open browser tabs, documents, calls, and Slack messages, storing everything as vector embeddings in a local, on-device database — nothing is uploaded to a cloud server — and feeds that context to Claude through a Model Context Protocol (MCP) connector. It uses Google's Gemini API specifically to generate the embeddings (a paid API call, though Google states it doesn't train on or retain that data) and Deepgram for meeting transcription, reporting 54% accuracy on the BEAM long-term-memory benchmark — a figure worth treating cautiously since there's no context given for how difficult that specific benchmark is. It's built by Shram Intelligence, Inc., published in July 2026, and pricing isn't published on the homepage, currently appearing to be a free Mac app.
| # | Tool | Best for | Price |
|---|---|---|---|
| 1 | Équipes entreprise et développeurs qui veulent déployer des agents IA de production fiables, avec gouvernance et sans dépendance à un seul fournisseur cloud | — | |
| 2 | Développeurs et équipes IA construisant chatbots, systèmes RAG, pipelines d'analyse de documents et traitement par lots, cherchant un coût d'infrastructure prévisible et une garantie de confidentialité | — | |
| 3 | Développeurs et organisations qui utilisent des LLM en production et veulent réduire la facture d'inférence sans changer leur code | — | |
| 4 | Développeurs et équipes qui construisent des agents IA ou des produits IA ayant besoin d'un accès de recherche structuré et unifié | — | |
| 5 | Créateurs solo, équipes internes, agences IA et responsables opérations voulant fiabiliser un agent IA avant/pendant la production | — | |
| 6 | Équipes en entreprise (finance, santé, retail, go-to-market) qui veulent une IA avec une mémoire partagée et auditable plutôt que des réponses individuelles incohérentes | — | |
| 7 | Équipes produit qui veulent intégrer des agents IA (Codex, Claude Code, Hermes) dans leur propre app sans construire elles-mêmes le sandboxing et l'orchestration | — | |
| 8 | Entrepreneurs matériels, makers et porteurs de produit qui veulent une première estimation de faisabilité avant d'investir dans le développement | — | |
| 9 | Courtiers en fret et entreprises de logistique gérant un volume important de réservations de camions | — | |
| 10 | Entreprises multi-sites (retail, restauration rapide, immobilier, santé, assurance, vente B2B) voulant unifier appels, recrutement, formation et coaching | — | |
| 11 | Opérateurs et dispatchers de flottes de camions cherchant à optimiser l'attribution des trajets et l'utilisation des chauffeurs | — | |
| 12 | Associations et ONG avec des années de documents non structurés (rapports, comptes-rendus, politiques) | — |
A toolkit for building AI "employees" that actually finish a job reliably — instead of a chatbot that just talks, Timbal helps a company wire up an AI agent that can look things up, take real actions in other systems, and behave consistently every time.
AI models on a phone-plan-style flat rate instead of a taxi meter — pay one predictable monthly price for a set number of requests across 40+ AI models, rather than watching a per-token bill climb with every word generated.
A price-comparison shopper for AI model calls — instead of always paying whatever your one chosen AI provider charges, Auriko checks prices across a dozen providers in real time and sends each request to whichever one is cheapest right now without changin
A search engine built for AI agents instead of humans — it hands back clean, structured answers instead of a page of blue links the agent then has to re-read and guess at.
A quality-control inspector for AI agents — before your AI assistant goes live and starts talking to real customers, AgentX tests it, points out exactly where it messes up, and suggests a fix, instead of you finding out from an angry customer.
A shared AI brain for your whole team — instead of everyone copy-pasting into their own private ChatGPT, PromptQL keeps one team-wide memory that gets smarter every time someone corrects it.
A single API that lets any app run AI coding agents like Claude Code or Codex in the background and get back real finished work — code, videos, documents — without the app's developers having to build the sandboxing and orchestration themselves.
Describe a hardware idea in one sentence and get back, in about two minutes, whether it's realistically buildable, what it would cost to manufacture, and a rough 3D concept — before you spend real money finding out the hard way.
A voice AI that calls dozens of truck drivers at once to negotiate a load's price, so a freight broker gets a booked truck in minutes instead of working the phone all day.
An AI 'employee' for frontline businesses like retail and fast-casual chains — it answers phones, screens job candidates, coaches staff, and flags when someone's about to quit, all from the same conversation data.
An AI dispatcher for trucking companies — it looks at every driver, load and rule at once and works out the best assignment plan, instead of a human juggling spreadsheets and gut instinct.
A nonprofit's grant reports, board minutes and policy documents pile up over the years and become impossible to search by keyword — DocuSmart lets staff just ask a question in plain English and get an answer pulled from that pile.
