franki-cli 0.1.1

AI-powered CLI for coding, pentesting, and SOC work — bring your own models

franki

AI-powered terminal assistant for developers and security professionals.

franki is an open source CLI that brings AI assistance directly to your terminal — for coding, CEH/pentesting, SOC analysis, and exam prep. It supports multiple free AI providers with automatic fallback when one hits a rate limit.

Install

pipx install franki-cli

Why pipx? franki is a CLI tool. pipx installs it in an isolated environment and puts the franki command on your PATH — no venv setup needed.

Install pipx first if you don't have it: sudo apt install pipx && pipx ensurepath

Or with pip in a virtual environment:

python3 -m venv .venv && source .venv/bin/activate
pip install franki-cli

After installing, run the setup wizard to add your API keys:

franki init

Quick start

franki                                    # launch interactive REPL
franki --version                          # print version
franki init                               # re-run setup wizard
franki config list                        # show config (keys masked)

Inside the REPL:

explain how TCP handshakes work           # plain message
@myfile.py refactor this function         # inject a file into context
!nmap -sV 192.168.1.1                     # run a shell command and analyse output
/skill pentest                            # switch skill
/help                                     # show all commands

Skills

Skill	Purpose
coding	Code generation, review, debugging
pentest	CEH-aligned recon, exploitation, reporting
soc	Log analysis, alert triage, incident response
ceh	CEH v13 exam prep and flashcard quizzes

Providers

franki works with multiple free AI APIs and automatically falls back when one hits a rate limit:

Provider	Free tier	Get key
Groq	Yes	groq.com
Google Gemini	Yes	aistudio.google.com
OpenRouter	Free models available	openrouter.ai
DelkaAI	Self-hosted option	—

Configure keys:

franki config set groq.api_key <key>
franki config set gemini.api_key <key>
franki config set openrouter.api_key <key>

Commands

Conversation

Command	Description
/clear	Clear conversation history
/compact	Summarise history to save context
/rewind	Undo the last exchange
/history	Show current session log
/context	Show model, memory, tokens, search status

Output

Command	Description
/export	Save session to Obsidian vault as markdown
/copy	Copy last AI response to clipboard
/note <text>	Save a timestamped finding note
/report	Generate a pentest or SOC report from the session
/search <query>	Web search via Tavily — injects results into context

Navigation

Command	Description
/skill <name>	Switch skill: coding / pentest / soc / ceh
/model <name>	Switch AI model
/scope <ip/cidr>	Set pentest target scope
/scope clear	Remove active scope

CEH / Security

Command	Description
/quiz	CEH v13 flashcard quiz mode
/mitre <behaviour>	Map a behaviour to MITRE ATT&CK
/payload <type>	Suggest payloads for an attack type
/tools <task>	Suggest the right tools for a task
/explain <tool>	Explain a tool, its flags, and usage

Memory

Command	Description
/remember <fact>	Save a fact to long-term memory
/memories	List all saved memory, scopes, skill usage, notes
/forget <id|all>	Remove a fact by id, or clear all memory

System

Command	Description
/connect	Show connection mode (delkaai / direct)
/connect delkaai	Switch to DelkaAI backend
/connect direct	Switch back to direct providers
/init	Re-run the API key setup wizard
/config	Open the config editor
/providers	Show provider status and configuration
/help	Show all commands
/quit	Exit (prompts to save session)

Auto-search

franki automatically runs a web search and injects the results before the AI responds when your message contains keywords like latest, current, today, news, or a CVE ID (CVE-XXXX-XXXXX).

Requires a Tavily API key (TAVILY_API_KEY) or a connected DelkaAI backend.

Long-term memory

franki remembers things across sessions:

/remember I use Python 3.11 and FastAPI
/remember my pentest lab is 10.10.10.0/24
/memories        # view everything stored
/forget 2        # remove entry #2

Stored facts, recent pentest scopes, skill usage, and notes are automatically injected into the system prompt at the start of every session.

Configuration

Config is stored at ~/.config/franki/config.json.

franki config list                    # show all config (keys masked)
franki config set groq.api_key <key>  # set a value
franki config get active_model        # read a value
franki config reset                   # reset to defaults

License

MIT — see LICENSE