ai-synapse 1.0.0

Free multi-provider AI SDK with terminal TUI, OpenAI-compatible API, and 27+ provider routing

AI Synapse

Free multi-provider AI SDK — drop-in OpenAI client, optional local server, and a terminal chat UI. Routes across 27+ providers with automatic failover, model caching, and intent-aware routing.

pip install ai-synapse          # SDK
pip install ai-synapse[tui]     # + terminal chat UI
pip install ai-synapse[all]    # SDK + TUI + server extras

---

Terminal UI (recommended)

ChatGPT-style chat in your terminal:

python -m ai_engine tui

Feature	How
Model picker	F2 — search, favorites (☆), paginated list
Provider	F3 — auto or pinned provider
Attach files	@filename fuzzy picker, or Attach button
Read file into prompt	/read + fuzzy picker
Slash commands	/help, /persona, /export, /defaults, …
Defaults	/defaults save · /defaults clear restore · shown in sidebar
Sidebar	Ctrl+B collapse · hidden scrollbar chat list
Copy	Ctrl+C / y — selection first, then last reply

Data lives under ~/.ai-engine/ (chats, preferences, personas).

---

Python SDK

No server required — call providers directly:

from ai_engine import OpenAI

client = OpenAI()
response = client.chat.completions.create(
    model="llama-3.3-70b-versatile",
    messages=[{"role": "user", "content": "Hello!"}],
    provider="groq",  # optional; auto-routes if omitted
)
print(response.choices[0].message.content)

Streaming:

stream = client.chat.completions.create(
    model="default",
    messages=[{"role": "user", "content": "Tell me a joke"}],
    stream=True,
)
for chunk in stream:
    piece = chunk.choices[0].delta.content or ""
    print(piece, end="", flush=True)

CLI:

ai-engine chat "Explain quantum tunneling"
ai-engine providers
ai-engine version

---

Local server (optional)

OpenAI-compatible HTTP API + web dashboard:

cp .env.example .env   # add free-tier API keys
pip install ai-synapse[server]
python -m ai_engine serve
# API: http://localhost:8000/v1/  ·  Docs: http://localhost:8000/docs

Use with the official OpenAI Python client:

from openai import OpenAI

client = OpenAI(base_url="http://localhost:8000/v1", api_key="dummy")

---

Providers

No key: Pollinations, Hermes, G4F proxies, OpenCode Zen, and more.

Free tier (signup): Groq, OpenRouter, Gemini, NVIDIA, Cerebras, Cloudflare, GitHub Models, Mistral, Cohere, Hugging Face, Vercel, Kilo.

Bring your own key: hcnsec, LLM7, PaxSenix, Ollama, self-hosted G4F.

Provider list and keys: see docs/collect_api.md. CDN config sync can refresh provider definitions at runtime (CDN_CONFIG_URL=default).

---

Configuration

cp .env.example .env

Set any keys you have — more keys mean more failover paths. The engine skips providers without credentials.

Variable	Purpose
GROQ_API_KEY	Groq free tier
OPENROUTER_API_KEY	OpenRouter free models
GEMINI_API_KEY	Google Gemini
GITHUB_TOKEN	GitHub Models
…	See .env.example

---

Project layout

Path	Role
ai_engine/	PyPI package — SDK, TUI, CLI
core/	Provider engine, cache, capabilities
server.py	Standalone FastAPI server (repo checkout)
docs/	API, deployment, security guides

---

Development

git clone https://github.com/mihir0209/AI_engine.git
cd AI_engine
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev,all]"
pytest tests/

---

Documentation

• API Reference
• User Guide
• Deployment
• Provider setup

---

License

MIT — see LICENSE.