Connecting…

Quick actions

↑ ↓ to move · Enter to choose · Esc to close · Ctrl+K to open

System resources

Live utilization on the Ollama server · refreshes every 5 seconds

Loading system telemetry…
CPU—

Loading processor details…

Memory—

Loading memory details…

Disk—

Loading filesystem details…

System—

Load average: —

GPU utilization

Loading GPU telemetry…

Loaded Ollama models

Checking Ollama…

GPU readings come from ROCm SMI when available. Model-to-device placement is shown when the server can match the Ollama runner to its GPU devices; Ollama reports model VRAM usage separately.

Saved messages

Keep useful answers and prompts easy to find across your chats.

Local usage analytics

Reply metrics from Ollama chats in your account. This report is computed on this device.

Model performance

Token and timing figures are returned by Ollama and stored with each account reply. Older replies may have only token and total-time data; temporary chats appear only while they remain open in this tab.

Conversation context

Keep a concise summary of this chat’s decisions, preferences, and open tasks. It is stored in your account and included only with this chat’s local Ollama requests.

Compare local models

Your prompt is sent to both selected Ollama models only. The current conversation is not included.

Choose two models and enter a prompt.

Both models run at the same time and may use substantial memory. Pick models that fit your server’s available RAM and VRAM.

Local model benchmark

Send the same standalone prompt to up to four installed Ollama models, one at a time.

Choose models 0 selected · up to 4
Choose models and enter a prompt.

Models run sequentially so only one benchmark request is active at a time. The prompt goes only to Ollama on this server; results stay in this tab until you download the CSV.

Ollama settings

Server-wide defaults · saved on this Ollama host

Loading server settings…

Reusable workflows

Load a plain-text SKILL.md from a public GitHub repository. Review its instructions before saving or enabling it. Supporting files and scripts are not imported.

Quick load:

Browse sources: Anthropic skills · Hugging Face skills. Imported instructions are treated as user-provided content and only affect chats when selected in Settings.

Scheduled tasks

Ollama runs these prompts on this server, even when this page is closed.

Your tasks

Library

Files created in Ollama replies and image tools. Saved in your account.

Plugins

Connect local tools and public services to your Ollama chats.

Plugins run only when you add them. Public GitHub repository details are fetched from GitHub when you send a message; chat content and credentials are never sent to GitHub. Adventure Land tokens are encrypted on this server; its read-only tools run only in chats where you enable them, and those tool requests go to adventure.land.

My models

Make focused Ollama assistants with their own instructions and base model.

Create a model

Choose an installed Ollama model and give this assistant a focused role.

Find a model

Search public GGUF models on Hugging Face and download them to this Ollama server.

Only public GGUF repositories are listed. Ollama chooses a recommended quantization automatically. Downloads use disk space on the Ollama server.

    Create a project

    Reference filesText, PDF and Word · up to 8 files

    Start a new chat

    Choose where this conversation belongs.

    Move chat

    Edits update the assistant message in this chat.
    Web browser

    Preview public HTTPS pages here. Some websites block embedding. Forms and pop-ups stay blocked. Avoid sign-in pages. Add notes below to give Ollama page context; notes stay in your browser until you send them.

    Enter to send · Shift+Enter for a new line

    Chats, attachments, projects, and settings are stored on this server in your account. Model requests are sent to the selected provider. Temporary audio and video uploads are removed after processing.

    ◉

    Ollama Chat

    Checking your account…

    Account