The ExecAI engine

Memory gives it context.
Agents put it to work.

ExecAI is a modular, memory-driven agent engine for business workflows, digital services and technical systems. It combines model inference, persistent knowledge, specialised agents and real tools in a runtime you configure and operate.

The system, at a glance

Separate responsibilities.
Connected capabilities.

An instance is the deployment boundary. Agents work inside it; memory, tools and model endpoints are attached through configuration. Select a component to explore its role.

EXECAI / ENGINE MAPDownload diagram

Scroll horizontally to explore the complete map.

ExecAI engine architecturePeople and applications enter a configured instance. Its access rules surround an agent runtime and message bus. Agents use configured memory stores, allowed tools and model backends through an admission queue. A separate instance can communicate through a Redis bridge. Memory sources and model endpoints can be shared only when explicitly configured. People · console · applications · API INSTANCE · ACCESS & PROCESS IDENTITY Agent runtimeIdentity · profile · context · permissions Agent registry & message bus Specialist agentsMessages & delegation Another instanceOptional Redis bridge Persistent memoryHybrid search · local embeddingsPrivate stores · read-only corpusDated facts & session distillation Tools & proceduresEquipment · files · web · visionSkills · MCP tool serversExplicit allowlists per profile Models & admission queueLocal: Ollama · llama.cppCloud: configured providersCapacity · priorities · cancellation Conceptual architecture · sources, permissions and connections are configured per deployment.
Deployment boundary

Instances define the perimeter.

An instance is a running ExecAI service, with its own configuration, process identity, sessions and access rules. Deploy separate instances for separate trust boundaries; an agent profile alone is not tenant isolation.

Conceptual architecture based on the project’s architecture, memory, agent and instance documentation. It does not imply automatic sharing or automatic isolation.

Persistent memory

Experience that outlives
a conversation.

An agent can write dated facts while it works and retrieve them later. Session distillation extracts durable knowledge, while updates correct existing records.

This is retrieval and knowledge maintenance, not retraining the underlying model. What an agent can recall depends on the sources configured for it.

RETRIEVE

Meaning and exact terms

Semantic vectors and BM25 lexical search combine through rank fusion. Local embeddings support on-site indexing; lexical search remains available when embeddings are unavailable.

ORGANISE

Private stores, deliberate sharing

Separate databases provide storage boundaries. Profiles can load multiple sources, including a common corpus in read-only mode. Tags organise records; they do not isolate tenants.

MAINTAIN

Facts with a history

Explicit writes, updates and session distillation keep knowledge useful. Review the stored facts and their sources instead of treating model recall as a guarantee.

Specialised agents

A mission. A profile.
A place in the team.

Identity blocks define instructions. Profiles select the model, tools, skills, memory, workspace and communication rights. The registry loads workers on demand.

ExecAI agent roster showing specialised profiles, models and configured capabilities.
Actual agent roster from the laboratory deployment · captured 17 November 2026

Coordinate work

Send asynchronous messages or delegate a task synchronously through the agent bus. Spawn rights are explicitly configured.

Connect expertise

Skills provide procedures. Tools provide access. MCP servers extend the tool catalogue without changing the agent’s mission.

Connect instances

An optional Redis bridge carries messages across instances. Each deployment still needs its own access and infrastructure policy.

Model freedom

The right model.
In the right place.

After auditing the workflow, we choose local models, cloud models or a combination. The decision starts with the task’s requirements and the model’s measured capability, then considers connectivity, privacy, hardware and operating cost.

Local backends include Ollama and llama.cpp; configured cloud backends include Qwen and DeepSeek. Each agent can have its own model. The shared admission queue coordinates capacity, priorities and cancellation.

DEPLOYMENT OPTIONS

Fully local

Models, embeddings, tools and knowledge run on your infrastructure. When all workflow dependencies are local, supervision and local alerting can continue without Internet. Prompts and operational data do not go to a model provider, and inference has no provider API fees. Hardware, electricity and maintenance still have a cost.

Deliberately hybrid

Keep routine or sensitive work local; use a cloud model for selected tasks that need more capability, where policy permits. The prompt, retrieved facts and tool results included in a cloud request are sent to that provider and incur its usage charges. Remote notifications also require connectivity.

Control at each layer

Set the perimeter.
Then set the permissions.

Security comes from configuration and infrastructure together. A prompt or an agent profile is not a substitute for process isolation.

IDENTITY

Admission & roles

OIDC authentication, admission lists and role-based access bound the users and agents available on a protected instance.

ACTION

Tools & modes

Tool allowlists and plan, edit or auto modes define the available actions and confirmation behavior.

PROCESS

Confinement

Configured tool sandboxes and operating-system service policies constrain paths, network access and process privileges.

VISIBILITY

Activity & usage

Persistent sessions, tool-call audit records and model usage logs support inspection. Their coverage depends on the enabled configuration.

The ExecAI console showing an agent conversation and current model information.
The console brings conversation, model configuration and operational visibility together · captured 17 November 2026
From architecture to installation

Build around
your constraints.

Start with the workflow, the data policy and the task. We will assess model capability and connectivity, then define a local or hybrid deployment and a pilot you can evaluate.