Open source Self-hosted Discord AI
Your bot.
Your infrastructure.
Shape the persona, select a provider, and choose which features your server needs. Freesona runs on your host, while requests to your selected AI provider are processed by that provider.
01 / A message, routed
Your setup, your request path.
01 / Discord
02 / Freesona
03 / Your provider
The diagram is simplified. Data handling depends on the provider and integrations you enable.
02 / Made of parts
One character. More than one setting.
Keep persona, memory, provider, and modules as separate choices. Change one without locking the rest in.
Shape a persona
/setpersona and save named presets.Keep memory distinct
Choose a provider
Configure optional modules
03 / Provider choice
Bring the model you prefer.
Built-in provider adapters include Gemini, OpenAI, Ollama, NVIDIA NIM, Azure AI Foundry, Groq, and OpenRouter. Freesona also supports custom OpenAI-compatible endpoints.
- Gemini
- OpenAI
- Ollama
- NVIDIA NIM
- Azure AI Foundry
- Groq
- OpenRouter
- Custom OpenAI-compatible
Provider availability, usage limits, and cost depend on the provider account, model, and hosting setup.
04 / Get started
Start with the quickstart.
Create your Discord application, configure your selected provider, then launch and test the bot.