Getting Started

Prerequisites

• Go 1.25.1 or higher
• Linux (sandbox via bubblewrap; auto-installs bwrap via apt/dnf/yum/pacman/apk if missing) or macOS (sandbox-exec)
• At least one LLM provider account (Copilot subscription, or an API key for OpenAI / Claude / Gemini / Nvidia)
• Optional: pdftotext (poppler-utils) for read_file to parse PDFs
• Optional: OPENAI_API_KEY to enable semantic search (text-embedding-3-small)

Install

git clone https://github.com/pardnchiu/Agenvoy.git
cd agenvoy
make build

make build compiles, embeds the latest git tag as projectVersion, and installs the binary to /usr/local/bin/agen.

Configure at least one provider

Agenvoy needs at least one LLM provider to operate:

agen model add

The interactive prompt walks through provider selection, model choice, and credential storage. Tokens land in the OS keychain (security on macOS, secret-tool on Linux, encrypted file fallback otherwise) under the fixed service name agenvoy.

The main config lives at ~/.config/agenvoy/config.json.

First run

# Create a named cli- session and switch the primary pointer to it
agen session new my-assistant

# Launch the full stack (TUI + Discord + Telegram + REST)
make app

Once the TUI is up, press i to open the Message input and submit with Enter (Shift+Enter inserts a newline on terminals that forward modifiers). Press c to open the Command ($) input. Tab toggles between Content and Logs in the main view; Ctrl+P opens the co-work dashboard (Sessions / Log / Pending three-panel).

For one-shot CLI usage:

make cli "summarize the latest changes in main.go"
make run "use playwright to open example.com and screenshot"

make cli confirms each non-read-only tool call; make run auto-approves everything.

Next steps

• Core Concepts — sessions, agent routing, the iteration loop, and three-pass tool dispatch
• Providers — supported LLM backends and the dispatcher model
• MCP Integration — plug in external tool servers
• CLI Reference — full command list

[!NOTE] This document was auto-generated by Claude after reading the full source code.