Elidia Agent CLI

Universal AI Agent for your terminal

Write code, search the web, manage files, run research, generate creative content — all from the command line.

---

What is Elidia Agent CLI?

Elidia Agent CLI is a powerful terminal-based AI agent that connects you to 30+ language models through a single interface. It goes beyond simple chat — it can execute tools, run multi-model consensus, conduct deep research, generate images/video/audio, and orchestrate complex workflows — all from your terminal.

Powered by the AiUtils Developer API, Elidia Agent CLI gives you access to models from OpenAI, Anthropic, Google, DeepSeek, Meta, Alibaba, and more — with a single API key.

Key Features

Multi-Model Intelligence

• 30+ models — Claude, GPT, Gemini, DeepSeek, Llama, Qwen, Mistral, and more
• Automatic model routing — picks the best model for each task based on complexity and cost
• Budget governance — session cost limits, pre-execution estimation, auto-downgrade to cheaper models

Smart Execution Modes

• Direct — fast single-model answers for simple questions
• Deep Think — extended reasoning with chain-of-thought for complex problems
• Consensus — runs 2-3 models in parallel, compares answers, synthesizes the best response
• Autonomous — multi-step agent loop with planning, tool use, and self-correction

Built-in Tools

• File system — read, write, search, and manage files
• Git — status, diff, log, commit — full git workflow from chat
• Terminal — execute shell commands with permission controls
• Web search — search the web and fetch page content
• MCP support — connect any MCP-compatible server for extended capabilities

Research Pipeline

• 5-stage YOYO pipeline — Decompose > Search > Analyze > Synthesize > Verify
• Multi-source — web search, academic papers, documentation
• Citation tracking — every claim linked to its source
• Export — Markdown, JSON, or plain text reports

Creative Generation

• Images — FLUX, DALL-E, Stable Diffusion (text-to-image)
• Video — Kling, Minimax (text-to-video, image-to-video)
• Speech — ElevenLabs, OpenAI TTS (text-to-speech)
• Music — Suno (text-to-music)

Workflow Engine

• YAML-defined pipelines — chain LLM calls, tool executions, and shell commands
• Conditions and loops — conditional branching and iteration
• Parallel execution — run steps concurrently
• Variable passing — capture outputs and feed into later steps

Memory & RAG

• Persistent memory — remembers context across sessions
• Vector search — sqlite-vec with 1024-dimensional embeddings
• Auto-compaction — keeps memory lean and relevant

Security & Permissions

• 4-tier permission system — AUTO, SESSION, EVERY_TIME, NEVER
• Progressive trust — permissions escalate based on usage patterns
• Audit logging — every tool execution is logged
• Keychain storage — API keys stored securely via OS keychain

---

Installation

From PyPI (recommended)

pip install elidia-agent-cli

From GitHub Release

Download the latest .whl file from Releases, then:

pip install elidia_agent_cli-0.1.0-py3-none-any.whl

Requirements

• Python 3.11 or higher
• An AiUtils Developer API key (get one here)

---

Quick Start

1. Authenticate

elidia auth login

You'll be prompted for your AiUtils Developer API key (ak-dev-...).

2. Start chatting

# Interactive REPL
elidia chat

# One-shot query
elidia chat "Explain async/await in Python"

# Specify a model
elidia chat --model claude-sonnet-5 "Review this function for bugs"

# Research mode
elidia chat --mode research "Compare PostgreSQL vs MongoDB for time-series data"

3. Check your balance

elidia auth status

---

REPL Commands

Once inside the interactive REPL (elidia chat), use slash commands:

Command	Description
/model [name]	Switch model or show current
/mode [mode]	Switch mode (chat, code, research, think, create)
/think [level]	Set thinking depth (minimal, low, medium, high, max)
/budget	Show session cost summary
/research <query>	Run deep research pipeline
/create image <prompt>	Generate an image
/create video <prompt>	Generate a video
/create speech <text>	Generate speech audio
/create music <prompt>	Generate music
/workflow <file.yaml>	Execute a YAML workflow
/tools [category]	List available tools
/mcp	Show MCP server status
/persona [name]	Switch agent persona
/memory search|save|forget	Manage persistent memory
/history [query]	Search chat history
/balance	Check DT balance
/cost	Session cost breakdown
/theme [name]	Switch UI theme
/cache [on|off]	Toggle response cache
/new	Start new session
/clear	Clear conversation
/help [command]	Show help

---

Configuration

Elidia Agent CLI stores configuration in ~/.elidia/config.toml:

[api]
base_url = "https://developer.aiutils.io/v1"
timeout_seconds = 120

[models]
default = "deepseek-chat"
code = "deepseek-chat"
reasoning = "deepseek-reasoner"
creative = "auto"

[permissions]
default_level = "session"

[budget]
session_limit_dt = 50.0
warn_threshold_dt = 40.0

[theme]
name = "default"

---

Available Models

Elidia Agent CLI supports 30+ models across multiple providers:

Provider	Models
Anthropic	Claude Sonnet 5, Claude Haiku 4.5
OpenAI	GPT-4o, GPT-4o-mini, o3, o4-mini
Google	Gemini 2.5 Pro, Gemini 2.5 Flash
DeepSeek	DeepSeek Chat, DeepSeek Reasoner
Meta	Llama 3.3 70B, Llama 4 Scout, Llama 4 Maverick
Alibaba	Qwen 2.5 72B, QwQ 32B
Mistral	Mistral Large, Codestral
Image	FLUX 1.1 Pro, FLUX Schnell, DALL-E 3, SD 3.5
Video	Kling v2, Minimax Video
Audio	ElevenLabs v2, OpenAI TTS, Suno v4

---

Workflow Example

Create a research.yaml file:

name: competitor-analysis
description: Research competitors and generate a summary

variables:
  topic: "AI coding assistants"

steps:
  - name: research
    type: llm
    prompt: "List the top 5 {{topic}} with their key features"
    output: findings
    model: deepseek-chat

  - name: analyze
    type: llm
    prompt: "Based on these findings, create a comparison table:\n{{findings}}"
    output: analysis
    model: claude-sonnet-5

  - name: save
    type: shell
    command: "echo '{{analysis}}' > competitor-report.md"

Run it:

elidia chat
> /workflow research.yaml

---

MCP Integration

Connect any MCP-compatible server for extended tool access:

# ~/.elidia/config.toml
[mcp.servers.filesystem]
command = "npx"
args = ["-y", "@modelcontextprotocol/server-filesystem", "/path/to/allowed/dir"]

[mcp.servers.github]
command = "npx"
args = ["-y", "@modelcontextprotocol/server-github"]
env = { GITHUB_TOKEN = "ghp_..." }

---

Local Models (Experimental)

For offline/free usage via Ollama:

# Install local model dependencies
pip install elidia-agent-cli[local]

# Use a local model
elidia chat --model ollama:llama3.2

---

API Key

Elidia Agent CLI uses the AiUtils Developer API. Get your API key at developer.aiutils.io.

Your key starts with ak-dev- and is stored securely in your OS keychain — never in plain text files.

---

Support

• Issues: GitHub Issues
• Email: support@aiutils.io

---

License

Proprietary software. Copyright 2026 Elidia Technology Pvt Ltd. All Rights Reserved.

See LICENSE for full terms.