Faam AI: an assistant that works inside your documents
Faam AI does more than chat. It reads the open document and calls the editor's tools to rewrite paragraphs, format text, insert tables, fix Excel formulas or add slides. You choose where the AI runs.
Three ways to use Faam AI
| Criteria | Faam AI Cloud | Your own API key | Local AI |
|---|---|---|---|
| Getting started | Sign in with a FaamOffice account | Paste your provider's API key | Install Ollama, LM Studio or llama.cpp |
| Cost | Credits from your account | Billed directly by the provider | Free (runs on your hardware) |
| Where data goes | Faam AI Cloud servers and the AI provider behind them | Straight to the provider you chose | Never leaves your machine (or your LAN) |
| Needs internet | Yes | Yes | No |
| Best for | Starting right away, no setup | Existing OpenAI, Claude, Gemini… accounts | Sensitive documents, working offline |
Faam AI Cloud
The fastest way to start: no API key, just a FaamOffice account. New accounts get free trial credits; your balance and usage history are on the Account page.
Create an account- 1Create a free FaamOffice account on this website.
- 2In the app, open Settings → AI models, choose Faam AI Cloud and click Sign in.
- 3Your browser opens a confirmation page. Check that the code matches the one in the app, then click Approve.
- 4Go back to the app and start working with Faam AI. You can sign the device out at any time from the Account page.
Use your own API key
Open Settings → AI models, choose a provider, paste your API key, pick a model and click Save. Requests go straight from your computer to the provider; FaamOffice never stores or relays your key through any server.
- OpenAI
- Anthropic Claude
- Google Gemini
- DeepSeek
- xAI Grok
- Mistral
- Groq
- OpenRouter
- Qwen
- Kimi
- GLM
- MiniMax
Run AI on your own machine
With local AI the model runs on your computer and document content never leaves it. Faam AI edits documents by calling tools, so choose a model with tool-calling support such as Qwen 3, Llama 3.1 or later, Mistral or GPT-OSS.
Ollama
Install Ollama from ollama.com, then download a model that supports tool calling:
ollama pull qwen3In FaamOffice: Settings → AI models → Ollama (local). Models are listed automatically from http://localhost:11434/v1; pick one and click Save.
LM Studio
Install LM Studio, download a model and start the local server with the lms command:
lms get qwen3
lms server startIn FaamOffice: Settings → AI models → LM Studio (local). Models are listed automatically from http://localhost:1234/v1; pick one and click Save.
llama.cpp
Run llama-server with a GGUF model (--jinja enables the chat template with tool calling):
llama-server -hf ggml-org/Qwen3-8B-GGUF --jinja --port 8080In FaamOffice: Settings → AI models → llama.cpp (local). Models are listed automatically from http://localhost:8080/v1; pick one and click Save.