Ecosystem map

The map is organized by the job a project does, not by repository age or deployment topology. Projects can be adopted independently; the arrows show integration seams, not mandatory dependencies.

flowchart TB
  subgraph design["Design & standards"]
    TAX[TAXOVELA]
    PAMP[PAMP]
  end
  subgraph foundation["Foundation"]
    PC[PRAXO-CORE]
    HV[Harmovela]
    CV[COORDIVELA]
  end
  subgraph runtime["Runtime & cognition"]
    PV[PRAXOVELA]
    MN[Mnemovela]
    NV[Noetivela]
  end
  subgraph work["Work & capabilities"]
    OD[ORCHADYN]
    MR[MODUREGIS]
    RV[RHEOVELA]
  end
  subgraph trust["Trust & knowledge"]
    EI[EIDOVELA]
    NM[NOMIVELA]
    AE[AEGIVELA]
    GN[GNOSIVELA]
    ON[ONTOVELA]
  end
  subgraph boundary["Boundary & assurance"]
    LM[LIMENORA]
    TK[TEKMOVELA]
    PR[PEIRAVELA]
    SD[SYNDOVELA]
  end
  subgraph human["Human & physical"]
    SB[SYMBIVELA]
    KN[KINETOVELA]
  end
  subgraph program["Program"]
    EA[Enterprise Autonomy]
  end

  PC --> PV
  PC --> NV
  PC --> MN
  PC --> HV
  PC --> GN
  PC --> AE
  PC --> LM
  PC --> RV
  EI --> NM
  NM --> AE
  GN --> ON
  OD --> MR
  MR --> RV
  RV --> SB
  TK --> PV
  TK --> RV
  PR --> KN
  TK --> KN
  SD --> MR
  TAX --> EA
  PAMP --> EA
  LM --> EA

Design and standards

What it answers: What should the system look like, and how do we know the design is sound?

• TAXOVELA turns requirements into architecture artifacts and validates them with deterministic gates.
• PAMP is a practical enterprise-architecture management platform for the portfolio around those designs.

These are design-time tools. They do not execute agents; they make the architecture around agents reviewable.

Foundation

What it answers: What do we build on, and how do independent pieces coordinate?

• PRAXO-CORE is the MIT-licensed agent framework: loop, primitives, adapters, policy, sandbox, plugins.
• Harmovela is the open protocol for asynchronous coordination: events, tasks, state, delegation, recovery, governance.
• COORDIVELA applies coordination to development work itself: leases, dependencies, conflicts, interventions.

The foundation is deliberately small. It provides seams, not a monolithic runtime.

Runtime and cognition

What it answers: Where does the agent actually run, what does it remember, and which model answers?

• PRAXOVELA is the local-first desktop agent runtime with a deny-by-default capability gate and an effect ledger.
• Mnemovela is the cognition runtime: typed memory, branch semantics, hybrid retrieval, governed forgetting.
• Noetivela is the inference fabric: model/endpoint catalog, inference contracts, and governed routing.

Work and capabilities

What it answers: What is the system allowed to do, what is the plan, and how does work survive failure?

• ORCHADYN compiles goals, state, capabilities, and constraints into verified, revisable plans.
• MODUREGIS is the capability control plane: publish, discover, authorize, invoke, audit.
• RHEOVELA turns governed capability plans into durable, recoverable process instances.

The typical flow is plan → govern → execute → verify, but each project is useful on its own.

Trust and knowledge

What it answers: Who is acting, what are they allowed to do, and what is true?

• EIDOVELA issues agent identity and workload attestation.
• NOMIVELA is the registry and namespace authority: identities, authority bindings, lifecycle epochs.
• AEGIVELA evaluates policy, attenuates delegated authority, and enforces decisions through PEPs.
• GNOSIVELA governs concepts, entities, claims, evidence, and conflicts, then projects scoped grounding views.
• ONTOVELA maintains an evidence-bearing, bitemporal operational world model and digital twin.

Identity and knowledge are separate concerns. A valid identity does not make a claim true; a grounded claim does not grant permission.

Boundary and assurance

What it answers: How does the system touch the outside world, and how do we know it still behaves correctly?

• LIMENORA is the governed north-south boundary for APIs, MCP servers, events, webhooks, and partner systems.
• TEKMOVELA turns intent into verification contracts, runs reproducible tests, and gates releases.
• PEIRAVELA runs possible-world experiments with provenance-complete simulation evidence.
• SYNDOVELA governs agent-skill composition and the supply chain around it.

Human and physical operations

What it answers: How do humans stay in control, and how does autonomy reach the physical world?

• SYMBIVELA is the human-agent collaboration workspace: goals, plan review, approvals, exceptions, intervention, outcomes, evidence.
• KINETOVELA is the physical-autonomy control plane for safe, observable robot-fleet missions.

Program

What it answers: How do the pieces compose into something end-to-end?

• Enterprise Autonomy is the public introduction and runnable-example program. It is the best place to see the products working together without reading every repository.

How to read the arrows

An arrow means “this project produces something the other can consume,” not “this project requires the other.” Examples:

• PRAXO-CORE exposes seams that NOETIVELA, MNEMOVELA, HARMOVELA, GNOSIVELA, AEGIVELA, LIMENORA, and RHEOVELA can fill.
• EIDOVELA issues identity, NOMIVELA records it, and AEGIVELA authorizes what the identity may do.
• ORCHADYN produces plans, MODUREGIS governs capabilities, and RHEOVELA executes and recovers process instances.

See the full relationship index for the edge list and the reasoning.