llmshot

A small bash command that talks to multiple LLM providers. One script, one interface. Switch providers with a flag.

Quick Example

$ llmshot -p openai -t "Give me a clever anagram I can impress my friends with"
Try this one:
A decimal point → I'm a dot in place
$

Features

• Multiple providers: OpenAI, Google (Gemini), Anthropic (Claude), and Ollama (local).
• Flexible prompt input: Inline text (-t), a file (-f), or stdin.
• Config file: API keys and default models in ~/.config/llmshot/llmshot.conf or via -e.
• Model override: Use -m to override the default model for the chosen provider.
• No extra runtime: Uses only bash, curl, and jq.

Prerequisites

You need:

• Bash (4+)
• curl
• jq

Installing prerequisites by platform

Linux

Debian / Ubuntu / Linux Mint

sudo apt update
sudo apt install curl jq

(Bash is usually already installed.)

Fedora / RHEL / CentOS / Rocky

sudo dnf install curl jq

Arch Linux

sudo pacman -S curl jq

openSUSE

sudo zypper install curl jq

Alpine

sudo apk add curl jq

macOS

Homebrew (recommended)

brew install curl jq

macOS ships with an old Bash; the script runs with /usr/bin/env bash (often Bash 3.2). For a newer Bash:

brew install bash

Then ensure your PATH uses the Homebrew bash before the system one if you rely on Bash 4+ features.

Windows

The script is written for Bash, so run it in a Bash environment:

Option A: WSL (Windows Subsystem for Linux)
Use your WSL distro (e.g. Ubuntu) and install as on Debian/Ubuntu:

sudo apt update && sudo apt install curl jq

Option B: Git for Windows (Git Bash)
Git for Windows includes Bash, curl, and often basic tools. Install jq separately:

• Download from jqlang/jq (e.g. jq-win64.exe), rename to jq.exe, and put it in your PATH, or
• If you have Chocolatey: choco install jq

Then run llmshot from Git Bash.

Option C: MSYS2 / Cygwin
Install the curl and jq packages in your MSYS2 or Cygwin environment and run the script from that shell.

---

Installation

One-command install (Linux / macOS / WSL)

curl -fsSL https://raw.githubusercontent.com/markabrahams/llmshot/main/install/install.sh | bash

This installs llmshot to ~/.local/bin (or /usr/local/bin if run as root). To use a custom directory:

curl -fsSL https://raw.githubusercontent.com/markabrahams/llmshot/main/install/install.sh | bash -s -- -d /usr/local/bin
# or
INSTALL_DIR=/usr/local/bin curl -fsSL https://raw.githubusercontent.com/markabrahams/llmshot/main/install/install.sh | bash

Ensure ~/.local/bin is in your PATH (e.g. add export PATH="$HOME/.local/bin:$PATH" to ~/.bashrc or ~/.profile).

Homebrew (macOS)

If you have a Homebrew tap that includes the formula (see scripts/llmshot.rb template in this repo):

brew install markabrahams/llmshot/llmshot

Manual

1. Clone or download this repo (or copy the llmshot script from bin/).

2. Make the script executable:

   chmod +x bin/llmshot

3. Put it on your PATH, e.g.:

   sudo cp bin/llmshot /usr/local/bin/
   # or
   mkdir -p ~/.local/bin && cp bin/llmshot ~/.local/bin/ && export PATH="$HOME/.local/bin:$PATH"

Configuration

Create a config file so llmshot can find your API keys and defaults.

Location (by default): ~/.config/llmshot/llmshot.conf

Override: use -e /path/to/file to point to another env/config file.

Config file format

Use the same variable names you would for environment variables. Example:

# OpenAI (optional if you only use other providers)
export OPENAI_API_KEY="sk-..."
export OPENAI_MODEL="gpt-4o"   # optional; script has a default

# Anthropic
export ANTHROPIC_API_KEY="sk-ant-..."
export ANTHROPIC_MODEL="claude-sonnet-4-5"   # optional

# Google (Gemini)
export GOOGLE_API_KEY="..."
export GOOGLE_MODEL="gemini-2.5-flash"   # optional

# Ollama (local; no API key)
export OLLAMA_MODEL="llama3.1"   # optional
export OLLAMA_URL="http://localhost:11434"   # optional

Only set the keys and variables for the providers you use. The script sources this file, so export is optional but keeps things clear.

Securing the config file

chmod 600 ~/.config/llmshot/llmshot.conf

Usage

Usage: llmshot -p <provider> [-m <model>] [-t <text>] [-f <file>] [-e <env_file>] [-u <url>]
  -p, --provider   Provider: openai, google, anthropic, ollama
  -m, --model      Override model from config (or use provider default if unset)
  -t, --text       Prompt text (takes precedence over -f and stdin)
  -f, --file       Prompt file (used if -t not given; else stdin)
  -e, --env        Environment/config file (optional; see above for default lookup)
  -u, --url        URL for ollama provider (default: http://localhost:11434)

Examples

Inline prompt (OpenAI):

./llmshot -p openai -t "Explain recursion in one sentence."

Prompt from file (Anthropic):

./llmshot -p anthropic -f ./prompt.txt

Stdin (e.g. pipe):

echo "Summarize the following: ..." | ./llmshot -p google

Ollama with custom URL and model:

./llmshot -p ollama -u http://192.168.1.10:11434 -m llama3.2 -t "Hello"

Override model for one call:

./llmshot -p openai -m gpt-4o-mini -t "Short joke about shells"

Use a specific config file:

./llmshot -p openai -e ~/work/llmshot.env -t "Hello"

Default models

If you don’t set a model in config or with -m, the script uses:

Provider	Default model
openai	gpt-5
google	gemini-2.5-flash
anthropic	claude-sonnet-4-5
ollama	llama3.1

(Update these in the script or in your config if your provider uses different model names.)

Troubleshooting

• "no environment file found"
  Create ~/.config/llmshot/llmshot.conf (or use -e <file>) with at least the API key for the provider you’re using.

• "Unknown provider"
  Use one of: openai, google, anthropic, ollama (case-insensitive).

• jq/curl not found
  Install jq and curl as in Prerequisites.

• Ollama connection errors
  Ensure Ollama is running (e.g. ollama serve) and that OLLAMA_URL (or -u) matches your server (default http://localhost:11434).

• API errors (401, 403, etc.)
  Check that the corresponding API key in your config is correct and has access to the requested model.

License

MIT License

Copyright (c) 2026 Mark Abrahams

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

Third-party services

This tool can interact with external AI providers such as OpenAI, Google, and Anthropic. This project is not affiliated with or endorsed by any of these providers. Users are responsible for complying with the respective API terms of service.