Ai

AI

Source

Eden's AI category covers the complete lifecycle of intelligent work: discover models and agents, run inference, route requests, give agents governed tools, reuse model state, and operate the result.

You can adopt each capability independently. The same identity, policy, telemetry, and evidence model connects them when you use them together.

Capabilities

CapabilityWhat it coversStart here
InferenceRun open and hosted models across local, private, cloud, and managed fleets.Inference
AI GatewayPut one governed API in front of models and providers.AI Gateway Governance
Models & providersConfigure endpoints, model eligibility, routing, fallback, limits, and cost.Model Routing And Orchestration
MarketplaceDiscover models, agent capabilities, tools, integrations, and eligible capacity.Marketplace
Agent FleetConnect agents and operate approved tools, context, workflows, identities, and runs as a fleet.Agent Fleet
RAGGround models and agents in permission-scoped context from connected systems.RAG
TrainingGovern datasets, training jobs, adapters, evaluations, checkpoints, and deployment.Planned
Shared AI memoryReuse permission-scoped model state near compatible inference runtimes.Shared Model State
AI operationsInspect usage, routing evidence, latency, cost, and recommendations.Analytics And Recommendations

How Eden implements AI

  • Runtime and Fleet coordinates model runtimes, fleets, placement, shared memory,

capacity, and execution.

  • Gateway supplies the gateway, identity, routing, policy, and provider boundary.
  • AI Orchestration supplies assistants, agents, skills, recipes, tools, and workflows.

No matter which capability you adopt, requests retain the same organization, identity, access, policy, telemetry, and audit context.

API surfaces

Help improve Eden Docs

Find something unclear or incomplete? Review the source and propose an update.

View on GitLab Updated July 27, 2026