Chatbot Ollama

About

Chatbot Ollama is an open source chat UI for Ollama.

This project is based on chatbot-ui by Mckay Wrigley.

If the image does not render, place the new screenshot at public/screenshots/screenshot-2025-09-dark.png (light version optional at public/screenshots/screenshot-2025-09-light.png).

Highlights (new)

• Blue theme polished for light and dark modes with consistent sidebars, headers, and overlays.
• Document upload (PDF, TXT, MD):
  • PDFs parsed server‑side for reliability; text/markdown read client‑side.
  • The full (truncated) content is attached invisibly to the model prompt — the UI stays clean.
  • 50 MB size cap, 100 page cap (PDF), 50k characters per attachment.
• Image upload: attach images and send them to vision‑capable models (base64). Non‑vision models simply ignore images.
• Code blocks: copy button, language label, optional line numbers, line‑wrap toggle, and “download as file”.
• Streaming & control: Stop Generation button, auto‑scroll toggle (pause/resume when you scroll up).
• Keyboard shortcuts:
  • Enter to send, Shift+Enter for newline
  • Esc to blur input
  • Ctrl/Cmd+L clears the composer
  • Arrow Up recalls the last user message
• Error handling: a consistent toast shows details, Copy details, and Retry.
• Conversation tools: Copy messages button copies the whole visible conversation (with role labels and any attached content).
• Sidebar UX: search inputs are theme‑aware, prompts list matches chat list styling, improved hover/selection.

Updates

Chatbot Ollama will be updated over time.

Next up

• Model management (pull/delete)
• Model info/details dialog

Docker

Build locally:

docker build -t chatbot-ollama .
docker run -p 3000:3000 chatbot-ollama

Pull from ghcr:

docker run -p 3000:3000 ghcr.io/ivanfioravanti/chatbot-ollama:main

Running Locally

1. Clone Repo

git clone https://github.com/ivanfioravanti/chatbot-ollama.git

2. Move to folder

cd chatbot-ollama

3. Install Dependencies

npm ci

4. Run Ollama server

Either via the cli:

ollama serve

or via the desktop client

5. Run App

npm run dev

6. Use It

You should be able to start chatting.

Tips

• Attach documents via the paper icon in the composer — the content is included for the model invisibly so the chat stays uncluttered.
• Attach an image with the camera icon — vision models will “see” it; others will ignore it.
• Use the clipboard icon in the sticky header to copy the current conversation.

Configuration

When deploying the application, the following environment variables can be set:

Environment Variable	Default value	Description
DEFAULT_MODEL	mistral:latest	The default model to use on new conversations
NEXT_PUBLIC_DEFAULT_SYSTEM_PROMPT	see here	The default system prompt to use on new conversations
NEXT_PUBLIC_DEFAULT_TEMPERATURE	1	The default temperature to use on new conversations

Contact

If you have any questions, feel free to reach out to me on X.