What is Fluent
Fluent helps AI assistants remember the parts of your life that matter, so their help gets more useful over time instead of starting from scratch in every new chat.
If you use AI for real decisions like what to cook, what to wear, or how to adjust your training, Fluent gives it the background it needs to be genuinely helpful.
The Problem
Most AI feels impressive at first, but not personal.
You explain your preferences, your routines, and your constraints. Then you open a new chat, switch to a different app, or come back a few days later and have to explain it all again. The result is advice that sounds polished but still feels generic.
What Fluent Does
Fluent gives your AI a lasting understanding of a few important parts of your life, beginning with Style, Meals, and Health.
That means your assistant can help using context that builds over time:
When you ask Claude to plan your meals for next week, Claude can use Fluent to pull in the context that answer depends on, including:
- What you like to eat and what you avoid
- What is in your kitchen inventory right now
- Your recipes and past feedback on meals you have cooked
- Your training program and whether you need more protein this week
When you ask Claude whether you should buy a jacket, Claude can use Fluent to review the relevant wardrobe context, including:
- Every item in your closet with fit profiles and photos
- Your style profile and what gaps exist in your wardrobe
- Whether the jacket overlaps with something you already own
Three Areas
Fluent currently focuses on three areas:
Style
A personal stylist that knows your entire closet. Import your wardrobe, get grounded purchase advice, identify gaps and redundancies, and build a profile of your aesthetic. Your assistant gives style guidance based on what you actually own and how you actually dress.
Meals
From planning to plating. Preferences, weekly meal plans, recipe management, grocery list generation, grocery ordering, kitchen inventory, and feedback loops. Your assistant knows your family's constraints, plans meals you'll actually cook, and orders the groceries.
Health
A coach that knows your program. Training preferences, goals, structured training blocks, session resolution, and block reviews. Your assistant adapts your training when life gets in the way — and works alongside Whoop, OURA, and the trackers you already use.
Two Deployment Tracks
Cloud Version
Fluent Cloud is coming soon. It is meant for people who want the Fluent experience without managing servers, storage, or updates themselves.
- Simpler setup for non-technical users
- Managed infrastructure and updates
- Same core Fluent experience, with less setup work
Fluent Self-Host (OSS)
The open-source version is available now at github.com/shaner-git/fluent-oss. You can run it on your laptop, a home server, or a VPS.
- Docker or bare Node.js
- SQLite + filesystem storage by default
- Optional experimental Postgres + S3 backend
- Bearer-token authentication
- Same core contract as Fluent Cloud
For Technical Users
Under the hood, Fluent runs as an MCP server. That means supported AI apps can connect to it as an external tool and use structured context from your Style, Meals, and Health data.
Both the self-hosted and cloud versions are designed around the same core contract: the same tools, the same resources, and the same payload shapes. Configurations and workflows written for one track are meant to carry cleanly to the other.
Architecture
+-------------------+ MCP +-------------------+
| Claude / Codex | <----------> | Fluent Server |
| (supported apps) | tools & | (Cloud or OSS) |
+-------------------+ resources +-------------------+
|
+------+------+
| Storage |
| D1/R2 |
| SQLite/FS |
| Postgres/S3|
+-------------+Fluent's shared core handles the MCP contract, onboarding, profile management, and domain logic. Runtime-specific concerns such as auth and storage adapters live at the edges.
Next Steps
- Cloud Version (Coming Soon) -- see what is planned for Fluent Cloud
- Quickstart (Self-Host) -- run Fluent on your own hardware
- Domains -- explore what each area can do