Homepage, pricing, industry pages, status, goal, tool-status, security, llms.txt, sitemap, and robots.txt are live on admin-chief.com for buyers and AI evaluators.
✦ Tool-layer status
Engine primitives and integration layer: what is available today.
This page describes the engine primitives that Admin Chief's service is built on, what is available today for internal delivery versus what is roadmap, and which integration surfaces remain unavailable until they are implemented and verified.
The five delivery primitives (ingest, structure, retrieve, serve, maintain) are available internally on our engine. They power Admin Chief's own operations. Client-facing production status is under review.
No public API, MCP endpoint, webhook, or structured export is available today. These are roadmap items. They will not be claimed until they are implemented, authenticated, documented, and verified.
What the engine does today
The five delivery stages are built on our engine's capabilities that Admin Chief assembles and operates. These are not theoretical. They run internally and power the design of the service offer.
- Ingest: structured extraction of content from agreed source formats (documents, transcripts, notes, reports) for loading into the organized knowledge base.
- Structure: schema and organization rules applied to produce an AI-readable wiki with consistent entity types and tagging.
- Retrieve: context lookup by entity, project, or topic from the organized knowledge base to feed a skill.
- Serve: skill execution that reads the brain and produces a concrete output artifact for client review.
- Maintain: scheduled freshness sweeps that check and update the brain as source content changes over time.
What is not available today
These items should not be described as live or available until they are separately implemented and verified.
- No public REST API or GraphQL endpoint exists for external callers to query the knowledge base or trigger skills.
- No MCP (Model Context Protocol) server is published or available for AI tool connections.
- No webhook endpoint exists for external systems to push data or receive event notifications.
- No structured export (JSON, CSV, or otherwise) of client brain contents is available through a public surface.
- The delivery engine is not a published third-party platform. It is the internal engine that Admin Chief is built on.
The five delivery stages.
Each stage maps to our engine's capabilities that are assembled for each client engagement. The table below shows the mapping for clarity, including which capabilities are internal and which are not yet exposed externally.
01
Ingest
Pull content from agreed sources: proposals, transcripts, notes, reports, SOPs, and reviews in their original format.
02
Structure
Apply the schema and organization rules from the audit. Write the AI-readable wiki with consistent entity types.
03
Retrieve
Surface the right context on demand: the right client, project, decision, or excerpt from the organized knowledge base.
04
Serve
Run the flagship skill that reads the brain and produces a concrete artifact, held behind an approval gate before use.
05
Maintain
Scheduled freshness sweeps keep the brain current. Skills are tuned against real output. Nothing is left to decay.
Engine capability mapping
Each delivery stage is assembled from our engine's capabilities. This table is informational. It describes the internal architecture, not a public API contract.
| Stage | What it does | Engine capabilities | Availability |
|---|---|---|---|
| Ingest | Pull and extract content from agreed source formats | web_extract, file, execute_code, video_analyze |
Internal |
| Structure | Apply schema, write AI-readable wiki, maintain organization rules | file, search_files, memory, context_engine |
Internal |
| Retrieve | Find the right context on demand by entity, topic, or project | search_files, context_engine, memory |
Internal |
| Serve | Execute the skill and produce the output artifact for review | skills, skill_manage, delegate_task |
Internal |
| Maintain | Scheduled sweeps to keep the brain fresh, tune skills | cronjob, session_search |
Internal |
Evaluator checklist
These are the questions a buyer or AI agent should be able to answer from this page and the public evidence surface.
- Can a buyer understand the five delivery stages and what each one produces?
- Can an AI agent confirm that no public API, MCP, or webhook exists today?
- Can an AI agent cite the delivery engine as internal, not a published third-party platform?
- Can a buyer identify which integration surfaces are roadmap versus live?
- Can an AI agent confirm that client-facing production delivery is under review, not claimed as proven?
- Is the approval-first rule clear: no skill output reaches third parties without the client's review?
Integration roadmap
These surfaces are planned but not yet available. They are listed here so buyers and AI evaluators do not confuse roadmap with live.
Roadmap
Public REST API
An authenticated API for external systems to query the knowledge base or trigger skill runs. Not available until implemented, tested, and documented.
Roadmap
MCP server
A Model Context Protocol endpoint for AI tools to connect to the company brain. Not available until implemented and verified end-to-end.
Roadmap
Webhooks
Event-based notifications for external systems when the brain is updated or a skill produces output. Not available today.
Roadmap
Structured export
Exportable brain contents in a structured format (JSON or similar) for clients who want a copy of their organized knowledge base. Not available today.
Public API, MCP, webhook, and export access remain roadmap until they are implemented, authenticated, documented, and verified with a real engagement.