Products
Product Areas Overview
Eden Substrate is organized around six customer-facing product areas. Each area describes a system or operational job customers already recognize. The products underneath those areas can be adopted independently and connected through one shared operating model.
All areas share the same organization, identity, endpoint, credential, policy, telemetry, and audit model.
For the detailed capability checklist behind these product areas, see the Product Feature Inventory.
Product areas
| Area | Primary job | Start here |
|---|---|---|
| AI | Discover, build, run, route, and govern models and agents. | AI |
| Storage | Run the storage primitives applications and AI depend on, in the infrastructure you control. | Storage |
| Data Gateway | Connect, route, compose, and act on data across governed systems. | Data Gateway |
| Modernization | Move and change production systems with validation, cutover, and recovery controls. | Modernization |
| Operations | Understand production, improve behavior, and operate fleets, cost, telemetry, and deployments. | Operations |
| Security | Protect AI, agents, data, identity, policy, approvals, evidence, and private connectivity. | Security |
Underlying products
| Product | Primary job | Typical users | Main docs |
|---|---|---|---|
| Eve | Govern endpoint and native gateway access to databases, APIs, tools, services, and model provider endpoints. | Platform engineers, security teams, application teams | Eve |
| Adam | Govern AI working against endpoints through the Eden LLM interface, tool calling, model routing, policy, and evidence. | Application teams, operators, analysts, agent builders | Adam |
| Crucible | Compare a primary endpoint with same-kind secondary endpoints under live traffic without changing the client response path. | Platform engineers, database owners, performance teams | Crucible |
| Exodus | Move live database workloads between source and target systems with validation and rollback. | Infrastructure teams, database owners, migration operators | Exodus Migrations |
Use one area independently or combine them. A database migration can use Storage, Data Gateway, Modernization, Operations, and Security without creating separate identity or telemetry systems. An AI workload can use AI, Storage, Data Gateway, Operations, and Security under the same organization and policy context.
Product implementation map
| Capability | What it provides | Documentation |
|---|---|---|
| Gateway | Governed access to endpoints, APIs, models, identity, policy, routing, and request evidence. | Gateway |
| Data plane | Routing, mirroring, protocol-aware execution, and controlled traffic changes. | Data plane |
| AI orchestration | Assistants, agents, tools, model routing, and governed AI workflows. | AI orchestration |
| Database migrations | Migration planning, movement, replay, validation, cutover, rollback, and recovery. | Database migrations |
| Runtime and fleet | Runtime fleets, model serving, placement, capacity, and execution coordination. | Runtime and fleet |
These capability names describe what teams can do. They are not dependent on which implementation or deployment boundary provides the capability.
Exodus is the migration product. It uses Eve endpoints and interlays to observe traffic, validate targets, move historical data, coordinate live traffic, and keep rollback available until the operator commits the migration.
Crucible is the comparison product. It attaches its own bounded runtime to an Eve interlay, dispatches eligible traffic to same-kind contenders, and records protocol-aware evidence without making comparison work part of the client response. Completion can select one contender behind the unchanged interlay. Crucible does not synchronize data and does not replace Exodus migration controls.
Shared concepts
The products share these concepts:
| Concept | Why it matters |
|---|---|
| Organization | Tenant boundary for users, agents, services, endpoints, workflows, fleets, and migrations. |
| Identity | The human, agent, service, runtime, or node allowed to act. |
| Endpoint | A named connection to a database, model provider, API, tool, service, or platform. |
| Access and policy | Controls what an identity can configure, call, route, move, or execute. |
| Runtime and route | Describe where eligible work can run and how a request reaches it. |
| Evidence | Safe records that explain actors, policy, route selection, execution, and outcomes. |
| Telemetry | Metrics, logs, traces, usage, cost, migration events, and fleet health. |
Which Product Should I Start With?
Start with Eve when you need a governed path to production systems. Eve is the right starting point for endpoint onboarding, RBAC, connection pooling, endpoint request routing, command policy, observability, and REST API usage.
Start with Adam when the user experience is conversational, agentic, model-driven, or needs AI-specific governance for model routes, budgets, tool calling, endpoint access through AI, or promptless evidence. Adam is the right starting point when licensed people, agents, or gateway API keys need to use models, discover tools, invoke endpoint operations, or work through governed AI workflows.
Start with Exodus when the business goal is a live database migration. Exodus is the right starting point when the source and target must run in parallel, traffic must be observed or mirrored, rollback must stay available, and operators need a staged migration runbook.
Start with Crucible when the question is whether two endpoint systems behave similarly under a real workload. Use it to collect comparison evidence before a migration, compatibility change, capacity decision, or provider evaluation.
Common Deployment Shape
A typical deployment has:
- Eve running as the API and gateway runtime.
- A backing control-plane store for organization, user, endpoint, and workflow state.
- One or more endpoint connections for databases, model providers, APIs, or tools.
- Optional interlays for protocol-level traffic such as Redis, PostgreSQL, MongoDB, or migration workflows.
- Observability export to the customer's metrics, logs, traces, or dashboard stack.