About VegaDūta

VegaDūta is an open-source, multi-tenant platform for building AI agents, wiring them into workflows with human approval, grounding them in your own data, and reaching them over the channels your users already use. It is self-hostable and released under the Business Source License 1.1 license.

Technology

Core: Java 21, Spring Boot, one Maven reactor of focused modules
Consoles: Next.js: tenant console and platform-admin console
Data: Postgres 16 with pgvector and Apache AGE; one schema per tenant
Cache and queues: Valkey; optional Temporal for durable workflows
Identity: Keycloak (OIDC), SSO, per-tenant isolation
Services: Python sidecar (ML), code-intel, sandboxed tool runner, MCP stdio runner

Standards

MCP: client and server (Streamable HTTP, SSE, stdio)
A2A: signed agent cards, JSON-RPC message/send and tasks
AG-UI: server-sent-event agent runs
OpenTelemetry: OTLP traces with gen_ai.* attributes
OIDC: Keycloak or any compliant identity provider
OpenAPI: 3.x and Swagger 2.0 import as tools

Documentation and dependencies

Documentation: install, configure, extend, architecture.
Java Agent SDK: governance for agents inside your own Java application.
TypeScript and Python clients: call the platform API from your code.
Plugins, apps and mobile: editors, browsers, the Android app and the MCP server, with their real status.

Project and governance

Source, issues and releases live on GitHub. Decision-making is described in GOVERNANCE.md; security reports follow SECURITY.md.

Operator

Maintained by: Cloudspace Learning and Development Private Limited
Contact: [email protected]