Features

• Terminal based
• Response highlights
• Chat history
• Download/Delete models from ollama server
• Simple
• Logging

Demo

Installation

Binary releases

You can download the pre-built binaries from the release page

crates.io

rtwo can be installed from crates.io

cargo install rtwo

Build from source

Prerequisites:

• Rust
• Cargo package manager.

Build

1. Clone the repo: git clone PATH_TO_REPO
2. Build: cargo build --release

This will produce an binary executable file at target/release/rtwo that you can copy to a directory in your $PATH.

Configuration

rtwo can be configured using a TOML configuration file. The file is located at:

• Linux : $HOME/.config/rtwo/rtwo.toml
• Mac : $HOME/Library/Application Support/rtwo/rtwo.toml
• Windows : {FOLDERID_RoamingAppData}\rtwo\config\rtwo.toml.

The default configuration is:

host = "localhost"
port = 11434
model = "llama3"
verbose = false
color = true
save = true
stream = false

• host: target host for ollama server
• port: target port for ollama server
• model: model to query
• verbose: enable/disable verbose output from responses (See Usage)
• color: enable/disable color output from responses
• save: enable/disable saving responses to DB ($HOME/.local/share/rtwo/rtwo.db)
• stream: enable/disable streaming response to terminal
  • NOTE: false is the recommended value. Setting this to true disables markdown output formatting. However, for slower ollama servers or larger models, it may be advantageous to set this value to true.

Usage

  -H, --host <HOST>
          Host address for ollama server. e.g.: localhost, 192.168.1.5, etc.
          Can be set in configuration file.
          Default: localhost

  -p, --port <PORT>
          Host port for ollama server. e.g.: 11434, 1776, etc.
          Can be set in configuration file.
          Default: 11434

  -m, --model <MODEL>
          Model name to query. e.g.: mistral, llama3:70b, etc.
          NOTE: If model is not available on HOST, rtwo will not automatically download the model to the HOST. Use "pull" [-P, --pull] to download the model to the HOST.
          Can be set in configuration file.
          Default: llama3

  -v, --verbose
          Enable verbose output. Prints: model, tokens in prompt, tokens in response, and time taken after response is rendered to user.
          Example:
          	* Model: llama3:70b
          	* Tokens in prompt: 23
          	* Tokens in response: 216
          	* Time taken: 27.174
          Can be set in configuration file.
          Default: false

  -c, --color
          Enable color output.
          Can be set in configuration file.
          Default: true

  -n, --nocolor
          Disable color output.
          Can be set in configuration file [color = false].

  -s, --save
          Save conversation for recall (places conversation in DB)
          Can be set in configuration file.
          Default: true

  -i, --incogneto
          Do NOT save conversation for recall.
          Can be set in configuration file [save = false].

  -b, --batch
          Do not stream llm output (wait for full response to generate before rendering to user).
          NOTE: This allows response formatting.
          Can be set in configuration file [stream = false].

  -S, --stream
          Stream llm output (Display response as it is rendered by host)
          NOTE: This disables response formatting.
          Can be set in configuration file.
          Default: false

  -l, --list
          List previous conversations

  -L, --listmodel
          List available models on ollama server (HOST:PORT)

  -r, --restore
          Select previous conversation from local storage and pick up where you left off. This restores the context
          from a saved conversation and prints the saved output.
          Interactive

  -d, --delete
          Delete previous conversations from local storage.
          NOTE: action is irreversible.
          Interactive

  -P, --pull <MODEL>
          Pull model to ollama server for use (downloads model on HOST). e.g.: llama3.

  -D, --delete-model <MODEL>
          Delete model from ollama server (deletes model on HOST). e.g.: llama2.

  -h, --help
          Print help (see a summary with '-h')

  -V, --version
          Print version

Donate

• BTC: bc1qvx8q2xxwesw22yvrftff89e79yh86s56y2p9x9
• XMR: 84t9GUWQVJSGxF8cbMtRBd67YDAHnTsrdWVStcdpiwcAcAnVy21U6RmLdwiQdbfsyu16UqZn6qj1gGheTMkHkYA4HbVN4zS

License

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.