Ollama integration
Use this guide to configure Ollama as a local AI provider for ReplyRabbit in FreeScout. Ollama lets ReplyRabbit request AI drafts from a model running on your own hardware.
Looking for the overview?
Overview
Ollama is a local provider option in ReplyRabbit. Local providers do not require an API key. Instead, ReplyRabbit needs a local server URL and a model name that exists in the Ollama server.
The default Ollama URL is http://localhost:11434. Use that URL unless your installed ReplyRabbit version documents another supported local setup.
Prerequisites
Before you start, make sure you have:
- A working FreeScout installation.
- ReplyRabbit installed and activated.
- ReplyRabbit Pro or Team, which includes local providers.
- Ollama installed and running where ReplyRabbit can reach it.
- A model installed in Ollama.
If you are still choosing between cloud and local providers, read Choosing an AI provider.
Local setup requirements
Ollama must be reachable by the PHP process running FreeScout. ReplyRabbit defaults Ollama to http://localhost:11434; use another URL only when your installed version documents that setup.
Do not assume arbitrary LAN hostnames or ports will be accepted. Use the default URL unless your installed ReplyRabbit build documents a different supported setup.
Connect Ollama
Start Ollama
Start Ollama and confirm the model you want to use is available in the local server.Open ReplyRabbit settings
In FreeScout, open the mailbox, go to Settings, and choose ReplyRabbit.Choose Ollama
In the AI Provider card, select Ollama (Local).Confirm the server URL
Usehttp://localhost:11434unless your installed ReplyRabbit version documents another supported Ollama URL.Choose or enter the model
Select a model option or use the custom model field. The model must exist in the running Ollama server.Save and test
Save settings, open a FreeScout conversation, and generate a draft. Review the result before using Ollama on a busy mailbox.
Configuration notes
- API key: not required for Ollama.
- Server URL: defaults to
http://localhost:11434. - Model: must match an installed Ollama model name.
- Quality: depends on the model and hardware you run.
- Budget: token usage can still be recorded by ReplyRabbit, but Ollama itself does not charge per token.
Use cases
- Keep AI generation on hardware you control.
- Test local model quality before using it for a production mailbox.
- Avoid per-token cloud provider billing for supported local workflows.
Troubleshooting
| Symptom | What to check |
|---|---|
| Invalid Ollama URL | Use http://localhost:11434 unless your installed version documents another supported local URL. |
| Model not found | Confirm the model name exists in Ollama and matches the ReplyRabbit setting. |
| Draft generation times out | Try a smaller model, check server resources, and confirm Ollama is running. |
| Output quality is weak | Test a stronger local model or switch the mailbox back to a cloud provider. |
| Ollama is disabled in settings | Confirm the mailbox is on ReplyRabbit Pro or Team. |
For broader fixes, see Troubleshooting ReplyRabbit.
Privacy and security notes
Ollama keeps generation local when the server and model run on your own hardware. That can help with data locality, but you still need to secure the machine, FreeScout, and any local services that can access support content.
If you expose Ollama beyond localhost, confirm that your ReplyRabbit version supports that setup and that the endpoint is not reachable by untrusted users.
Next steps
- Review the Ollama integration for FreeScout page when you need the workflow and buying overview.
- Compare local and cloud tradeoffs in Choosing an AI provider.
- Continue with How to set up ReplyRabbit if company context, examples, or budgets still need attention.
FAQ
Does Ollama require an API key?
No. ReplyRabbit uses an Ollama server URL and model name instead of a cloud provider API key.
What Ollama URL should I use?
Use http://localhost:11434 unless your installed ReplyRabbit version documents another supported local setup.
Can I use any model name?
The model name must match a model available in your running Ollama server.
Is Ollama available on the Free tier?
No. Local providers are available on ReplyRabbit Pro and Team.