Vibe Coding Discoveromnicoreagent
View on GitHubOpen Python agent harness for production AI apps: tools, MCP, memory, workspace, telemetry, subagents, background tasks, and OmniServe APIs.
OmniCoreAgent is an open Python runtime for building production AI agents: tools, MCP clients, memory, workspace, subagents, background tasks, sandboxed execution, governance/budget policies and tracing. Ships OmniServe to expose agents over REST/SSE.
Use Cases
build production AI agents with tools and memoryconnect agents to MCP serverssandboxed code/command execution for agentsdurable runs that resume after crash or approvalcost and token budget enforcement per agent/appgovernance policies (allow/ask/deny) over capabilitiesbackground and scheduled agent tasksspawn subagents under shared policy and budgetsend-to-end run tracing with OTLP exportserve agents as REST/SSE APIs via OmniServesession persistence in Redis/Postgres/MongoDBmulti-provider LLM access through one model config
Built With
- Language
- Python
- Frameworks
- LiteLLM · FastAPI · Pydantic · MCP SDK · OpenTelemetry · Redis · Docker · E2B · Modal · Daytona · Vercel Sandbox · PostgreSQL · MongoDB · boto3/S3 · LangSmith · Opik
Tags
agent-harness · multi-agent · subagents · mcp · memory · telemetry · durable-runs · sandboxing · governance · budgets · background-tasks · guardrails · observability · orchestration · python · fastapi