Agentic LLM router for classification and decision making.
Project description
Agentic Router
Agentic Router is a domain-agnostic, zero-shot classification and routing engine designed for modern LLM applications.
Instead of writing rigid if/else statements or hardcoding API logic, Agentic Router allows you to define complex routing topologies and classification guardrails using simple Python dictionaries and arrays. It dynamically compiles these into robust framework-agnostic prompts to intelligently route user queries to the correct database or pipeline.
Features
- 🛣️ Dynamic Topologies: Route natural language queries across various workflows (e.g., Vector DB vs Graph DB vs SQL) by passing a simple dictionary of rules.
- 🛡️ Custom Classification: Inject arbitrary categories and guardrails (e.g., PII detection, Sentiment Analysis) into the agent's decision-making pipeline on the fly.
- 🔌 Model Agnostic: Built on top of LlamaIndex. Inject an OpenAI, Gemini, or even a local Ollama model directly into the constructor.
Installation
pip install agentic-router
Quick Start
1. Dynamic Query Routing
Define your project's unique routing rules, and the Agent will dynamically strictly enforce them:
from agentic_router import AgenticRouter
from llama_index.llms.openai import OpenAI
llm = OpenAI(model="gpt-4o")
router = AgenticRouter(llm=llm)
# A Customer Support topology
support_topology = {
"Tech_Support": "User is asking about a software bug or failure.",
"Billing": "User is asking about refunds or credit cards."
}
route = router.route_query(
query="I was charged twice on my visa.",
topology=support_topology,
fallback="Tech_Support"
)
print(route) # Output: Billing
2. Custom Classification & Guardrails
Classify inputs on the fly strictly adhering to your defined arrays:
# A tabletop RPG setting
rpg_vibes = ["Ominous", "Peaceful", "Combat", "Tavern"]
query = "The floorboards creak as three figures step out of the shadows with drawn daggers."
vibe = router.classify_vibe(query=query, valid_vibes=rpg_vibes)
print(vibe) # Output: Ominous
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file agentic_router-0.1.0.tar.gz.
File metadata
- Download URL: agentic_router-0.1.0.tar.gz
- Upload date:
- Size: 4.7 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
4167a0a4483d6ec60f84e2e118b073695a7f383762057778ed15c32ec56fc7d2
|
|
| MD5 |
90e031744369fa9e38afcfe23935c725
|
|
| BLAKE2b-256 |
40c89b8d849998bda9154e88221b1359113c3302e0cde8d7dfdb45e1143c4615
|
File details
Details for the file agentic_router-0.1.0-py3-none-any.whl.
File metadata
- Download URL: agentic_router-0.1.0-py3-none-any.whl
- Upload date:
- Size: 4.4 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
f412798c8f53668d374a67cf20a3feda07d483a52fe9388d25a5115a1d2db883
|
|
| MD5 |
53898cba818201fba0c5e154ed2af6bf
|
|
| BLAKE2b-256 |
fd9cd3aeb7d39de41794b68788bb10ce220f5d19fd60cafb7f41f9078f14a895
|