Skip to main content

​
API Keys

Fliiq needs at least one LLM API key. Add them to ~/.fliiq/.env:
.env
# LLM providers (at least one required)
ANTHROPIC_API_KEY=sk-ant-...
OPENAI_API_KEY=sk-...
GEMINI_API_KEY=...
Priority order: Anthropic > OpenAI > Gemini. If your primary provider fails, Fliiq automatically falls back to the next available one. To change the default, see Provider Switching below. For best results with Anthropic (recommended provider):
ModelBest forNotes
Claude Sonnet 4.5Most tasksBest balance of speed and capability. Recommended default.
Claude Opus 4.6Heavy computationComplex refactors, deep research, multi-step reasoning. Overkill for most use cases — Sonnet handles the vast majority of tasks well.
Fliiq works with any model your provider supports. Sonnet 4.5 is the sweet spot for daily use.

​
Provider Switching

When multiple API keys are set, Fliiq uses the default priority order (Anthropic > OpenAI > Gemini). Override this per-command with the --provider flag: 
fliiq --provider openai               # REPL with OpenAI/GPT-4o
fliiq run "task" --provider gemini     # Single-shot with Gemini
fliiq tui --provider anthropic         # TUI with Claude
For a persistent default, set FLIIQ_PROVIDER in your .env:
.env
FLIIQ_PROVIDER=openai
Resolution order:
  1. --provider flag (highest priority)
  2. FLIIQ_PROVIDER environment variable
  3. First API key found (Anthropic > OpenAI > Gemini)
See CLI Reference for the full list of commands that accept --provider.

​
Model Selection

Switch models per-command with --model (or -m): 
fliiq run "build a Jira skill" --model opus-4.6    # Use Opus for complex tasks
fliiq run "check my inbox" --model sonnet-4.5       # Use Sonnet for daily tasks
fliiq run "task" --model gpt-4.1                    # Use OpenAI GPT-4.1
View all available aliases with: 
fliiq models

​
Built-in Aliases

Default aliases shipped with fliiq init:
AnthropicOpenAIGemini
opus-4.6gpt-4.1gemini-2.5-pro
sonnet-4.5gpt-4.1-minigemini-2.5-flash
sonnet-4.0gpt-4o
haiku-4.5gpt-4o-mini
o3
o3-mini
o4-mini

​
Custom Aliases

Aliases are defined in ~/.fliiq/models.yaml (created by fliiq init). Add your own for Ollama, Mistral, DeepSeek, or any other provider: 
# ~/.fliiq/models.yaml
aliases:
  # ... default aliases above ...

  # Custom aliases — add yours here
  mistral-large: mistral-large-latest
  mistral-small: mistral-small-latest
  mistral-codestral: codestral-latest
  deepseek-v3: deepseek-chat
  deepseek-coder: deepseek-coder
  llama-3.2: llama3.2:latest
  qwen-2.5: qwen2.5:latest
Then use them like any built-in alias: 
fliiq run "task" --model mistral-large --provider openai
Any string not found in aliases is passed through as-is to the provider API.

​
Persistent Model Override

Override the default model for a provider via environment variable in your .env:
.env
ANTHROPIC_MODEL=claude-opus-4-6
OPENAI_MODEL=gpt-4.1
GEMINI_MODEL=gemini-2.5-flash
Resolution order: --model flag (highest) > env var (ANTHROPIC_MODEL, etc.) > hardcoded default.

​
Self-Hosted LLMs

Fliiq can connect to any OpenAI-compatible server — Ollama, vLLM, llama.cpp, LM Studio, LocalAI, or any other server that implements the OpenAI API. Set OPENAI_BASE_URL to point Fliiq at your local server:
.env
OPENAI_API_KEY=not-needed
OPENAI_BASE_URL=http://localhost:11434/v1
OPENAI_MODEL=llama3.2
OPENAI_API_KEY must still be set (most local servers ignore it — any non-empty value works).

​
Quick Start with Ollama

# 1. Install Ollama (https://ollama.com) and start it
ollama serve

# 2. Pull a model
ollama pull llama3.2
Add to ~/.fliiq/.env:
.env
OPENAI_API_KEY=not-needed
OPENAI_BASE_URL=http://localhost:11434/v1
OPENAI_MODEL=llama3.2

# 3. Run Fliiq with your local model
fliiq --provider openai
If you also have a cloud API key set, switch between local and cloud with --provider: 
fliiq --provider openai      # Local Ollama
fliiq --provider anthropic   # Cloud Claude

​
Anthropic-Compatible Proxies

If you use a proxy that implements the Anthropic API, set ANTHROPIC_BASE_URL:
.env
ANTHROPIC_BASE_URL=https://your-proxy.example.com
Self-hosted models vary in capability. Fliiq’s agent loop works best with models that support tool/function calling. Most large models (Llama 3.2+, Mistral, Qwen) support it via the OpenAI-compatible API.

​
Integration Credentials

Optional — add these when you need the corresponding skills:

​
Google (OAuth)

.env
GOOGLE_CLIENT_ID=your-client-id.apps.googleusercontent.com
GOOGLE_CLIENT_SECRET=GOCSPX-your-client-secret
Required for fliiq google auth — the OAuth flow that authorizes Fliiq to access Gmail, Google Calendar, Drive, Sheets, Docs, and Slides. See Google Integration for full setup. Used by: send_email, receive_emails, mark_email_read, archive_email, delete_email, google_calendar, google_drive, google_sheets, google_docs, google_slides

​
Gmail — Fliiq’s Bot Email

.env
FLIIQ_GMAIL_ADDRESS=my-fliiq@gmail.com
FLIIQ_GMAIL_APP_PASSWORD=xxxx-xxxx-xxxx-xxxx
Gives Fliiq its own email address as a channel. The app password is optional if you’ve authorized this account via OAuth instead.

​
Gmail — App Password (legacy)

.env
GMAIL_ADDRESS=you@gmail.com
GMAIL_APP_PASSWORD=xxxx-xxxx-xxxx-xxxx
Fallback for email skills when OAuth is not configured. Requires a Google App Password. We recommend OAuth instead — it’s more secure and unlocks Calendar access.

​
Twilio (SMS)

.env
TWILIO_ACCOUNT_SID=AC...
TWILIO_AUTH_TOKEN=...
TWILIO_PHONE_NUMBER=+1234567890
Get these from the Twilio Console. Used by: send_sms, receive_sms

​
Telegram

.env
TELEGRAM_BOT_TOKEN=123456:ABC-DEF...
TELEGRAM_ALLOWED_CHAT_IDS=12345678,87654321
Create a bot via @BotFather on Telegram. TELEGRAM_ALLOWED_CHAT_IDS restricts which chats the bot responds to. Used by: send_telegram, send_telegram_audio, Telegram real-time listener (daemon)

​
MiniMax (Text-to-Speech)

.env
MINIMAX_API_KEY=...
MINIMAX_GROUP_ID=...
Get these from the MiniMax Platform. The Group ID is found in your account settings. Used by: text_to_speech
.env
BRAVE_API_KEY=BSA...
Get a free key from Brave Search API. Used by: web_search

​
Spotify

.env
SPOTIFY_CLIENT_ID=...
SPOTIFY_CLIENT_SECRET=...
SPOTIFY_ACCESS_TOKEN=...
Create an app at the Spotify Developer Dashboard. Used by: spotify

​
Directory Structure

​
Global (~/.fliiq/)

Created by fliiq init. Used from any terminal.
PathPurpose
.envAPI keys and credentials
user.yamlUser profile — name, labeled email accounts (details)
models.yamlModel aliases for --model flag (details)
google_tokens.jsonOAuth tokens for authorized Google accounts (auto-managed)
memory/Persistent memory files
memory/MEMORY.mdCurated memory (always loaded into prompt)
audit/Audit trails from every agent run
skills/User-generated skills (available globally)

​
Project (.fliiq/) — optional

Created by fliiq init --project. Overrides global for this project.
PathPurpose
SOUL.mdAgent personality overrides
playbooks/Custom domain playbooks
mcp.jsonMCP server connections
memory/Project-specific memory
audit/Project-specific audit trails
jobs/Scheduled job definitions (YAML)
skills/Project-specific skills

​
Bundled (inside the package)

Ships with Fliiq. Cannot be modified directly.
PathPurpose
data/skills/core/33 bundled skills
data/soul/SOUL.mdDefault agent personality
playbooks/coding.mdDefault coding playbook

​
Config Resolution

When Fliiq looks for a resource (skills, SOUL.md, playbooks, memory):
  1. Project .fliiq/ — checked first
  2. Global ~/.fliiq/ — fallback
  3. Bundled defaults — final fallback
This lets you customize per-project without affecting your global setup, and ensures sensible defaults when nothing is configured.