Nexus

Enterprise-ready agent framework. Describe agents in config. Wire who is calling and where data lives at run time. Call run().

No global LLM settings. No shared agent singleton. Built for multi-tenant SaaS apps.

Beta — under active development. See NEXUS_AGENT_PRD.md for the design spec.

Why Nexus?

• SaaS-native — tenant, user, API keys, and storage passed per request
• Context-first (RCS) — compresses old tool results inline; no extra LLM calls
• Config-driven — YAML manifests for teams; Python for wiring and tools
• Provider-agnostic — bring your own model; pluggable storage
• Voice — gRPC media servers + WebSocket browser UI on the same agent core

Install

Published on PyPI as nexus-enterprise-agent. Import as nexus.

pip install nexus-enterprise-agent
# or
uv add nexus-enterprise-agent

pip install "nexus-enterprise-agent[sqlite,litellm,fastapi,realtime,grpc]"

Contributors:

git clone https://github.com/gowrav-vishwakarma/nexus-enterprise-agent.git
cd nexus-enterprise-agent
uv sync --extra sqlite --extra file
cp .env.example .env   # set your LLM API key

Text agents (YAML)

Three files: manifest, prompts module, short runner script.

uv run python examples/orchestration/run_team.py "Analyze Q4 revenue"

• Manifest: examples/orchestration/research_team.yaml
• Prompts: examples/orchestration/research_team_prompts.py

Python API (same agents, built in code):

from nexus import AgentRunner, RunContext
result = await AgentRunner(config=agent_config, tool_registry=registry, run_context=ctx).run("Hello")

Walkthrough: docs/getting-started.md (YAML) · docs/getting-started-python.md (Python)

Voice agents (gRPC + Voice Lab)

Cascaded voice: VAD → STT → LLM → TTS over gRPC media servers. Browser connects via WebSocket.

./scripts/run_voice_lab.sh

Opens http://localhost:8787 — allow microphone, click the mic button.

• Manifest: examples/orchestration/voice_grpc.yaml
• Media servers: examples/servers.yaml
• UI: examples/voice_lab.py

Python API (same manifest, custom transport):

from nexus.realtime.runtime import RealtimeRuntime
from nexus.realtime import RealtimeSession
from nexus.realtime.transport.websocket import WebSocketTransport

runtime = RealtimeRuntime.from_manifest(manifest, run_context=ctx)
pipeline = runtime.build_pipeline("voice_grpc")
session = RealtimeSession(pipeline, WebSocketTransport(websocket), session_id=sid)
await session.run_audio()

Guide: docs/guides/voice-lab.md · docs/guides/model-servers.md

Learn more

Full docs: docs/index.md