Features

Built for how enterprises actually work

Every feature in Aegis was designed for organizations that need air-tight security, compliance, and accuracy — not just convenience.

Connectors

Connect Every Data Source

Pre-built, production-grade connectors for the tools your teams already use. Incremental sync keeps everything up to date without full re-ingestion.

  • Confluence, SharePoint, Google Drive, Notion
  • Slack, Microsoft Teams, Jira, GitHub
  • S3, local file system, database connectors
  • Incremental sync with change detection
  • Custom connector API for proprietary sources
Confluence
SharePoint
Google Drive
Slack
GitHub
Jira

Pipeline

Production-Grade RAG Pipeline

A battle-tested retrieval-augmented generation pipeline with configurable chunking, hybrid search, reranking, and full observability.

  • Configurable chunking strategies (semantic, fixed, recursive)
  • Hybrid search: dense vectors + sparse BM25
  • Cross-encoder reranking for precision
  • Configurable embedding models (local or API)
  • Pipeline observability and metrics
1pipeline:
2 chunker: class="text-emerald-400">"semantic"
3 chunk_size: 512
4 chunk_overlap: 64
5 embedder: class="text-emerald-400">"local/bge-base"
6 search:
7 strategy: class="text-emerald-400">"hybrid"
8 top_k: 20
9 rerank: true
10 reranker: class="text-emerald-400">"cross-encoder"
11 final_k: 5

Access Control

Document-Level ACL

Permissions from your source systems are automatically synced and enforced at query time. Users only see answers from documents they're authorized to access.

  • Automatic permission sync from source connectors
  • Real-time ACL enforcement during retrieval
  • Group and user-level granularity
  • Permission audit trail and reporting
  • Zero-trust architecture — deny by default
1# Query-time ACL enforcement
2results = vectorstore.search(
3 query=class="text-emerald-400">"retention policy",
4 top_k=10,
5 acl_filter={
6 class="text-emerald-400">"user": class="text-emerald-400">"alice@corp.com",
7 class="text-emerald-400">"groups": [class="text-emerald-400">"engineering", class="text-emerald-400">"all"]
8 }
9)
10# Only returns documents Alice can access

AI Agent

Conversational AI with Citations

A RAG-powered agent that synthesizes answers from your knowledge base, always citing the specific documents and passages used.

  • Multi-turn conversation with session history
  • Source citations with document links and snippets
  • Configurable LLM backend (local or API)
  • System prompt customization per deployment
  • Streaming responses for low-latency UX
1agent:
2 llm: class="text-emerald-400">"local/mistral-7b"
3 temperature: 0.1
4 max_tokens: 1024
5 system_prompt: |
6 You are a knowledge assistant.
7 Always cite your sources.
8 citations: true
9 streaming: true
10 history: true

Integrations

Bring Your Own Stack

Aegis integrates with your existing infrastructure. Choose your LLM, vector store, identity provider, and deployment method.

  • LLM: OpenAI, Anthropic, local models (vLLM, Ollama)
  • Vector store: Qdrant, Milvus, Weaviate, pgvector
  • Auth: OIDC, SAML, LDAP, trusted proxy headers
  • Deploy: Docker, Kubernetes, bare metal
  • Monitoring: Prometheus, OpenTelemetry
LLMs
Vector DBs
SSO / OIDC
Kubernetes
Prometheus
OpenTelemetry

Plugins

Extend with Custom Plugins

Aegis's plugin architecture lets you add custom connectors, parsers, embedders, vector stores, and adapters — all via a clean Python interface and entry points.

  • Decorator-based plugin registration
  • Entry point discovery for third-party packages
  • Type-safe ABC interfaces for every plugin category
  • Hot-reload support during development
  • Dependency injection for clean wiring
1from aegis.core.registry import register_plugin
2from aegis.core.abc import BaseConnector
3
4@register_plugin(class="text-emerald-400">"connector", class="text-emerald-400">"custom-db")
5class CustomDBConnector(BaseConnector):
6 async def fetch(self, **kwargs):
7 # Your custom logic here
8 docs = await self.db.query(...)
9 return [Document(d) for d in docs]

See Aegis in Action

Schedule a personalized demo and see how Aegis transforms your team's knowledge into a secure, searchable resource.

Request a Demo