XCORE™ Tactical AI Analyst
XCORE™ Tactical AI Analyst

XCORE™ Tactical AI Analyst

XCORE™ — Local Off-Grid AI Analyst for XTOC

  • XCORE AI

XCORE™ is a local-first tactical AI analyst for real operations. It runs on your own Windows laptop, reads your local XTOC/XCOM data, and turns it into fast, grounded outputs: AO reviews, ADS-B pattern checks, 24h SITREPs, anomaly scans, and XTOC COM packet drafts.

Why it is a force multiplier:

  • Cloudless speed: on-device CPU GGUF models. No remote LLM dependency.
  • Grounded on your AO: answers are built from your local logs, roster, packets, SITREPs, and cached ADS-B heard aircraft (when available).
  • One-click workflows: AO review, aircraft frequency ranking, SITREP assembly, anomaly scan, and packet drafting.
  • Off-grid ask/reply: move AI requests and answers over any transport using AIASK/AIREPLY packet lines (field -> TOC -> field).
When the internet disappears, your analyst stays online.

Buy it here: XCORE™ – Tactical AI Analyst + XTOC Integration (Windows) – MKME Store

What XCORE™ is

XCORE™ is a standalone Windows desktop app with chat threads, curated Model Library, packet tooling, and bridge service endpoints for XTOC/XCOM integration.
  • Local data grounding: indexes selected XTOC/XCOM folders and optional snapshots to answer from your own records.
  • XTOC integration: invoke analysis from the XTOC Intel page using Run through XCORE™.
  • XCOM integration: run prompts from the field, generate AIASK packet lines, and decode AIREPLY replies locally.
  • LAN optional: run XCORE™ on a TOC laptop and let trusted clients query it over a controlled network.
If your XTOC/XCOM data is on the machine, XCORE™ can reason on it locally.
XCORE AI

What it can do (today)

  • Review your current AO: posture, concerns, highlights, and people-in-AO summaries grounded in recent XTOC records.
  • ADS-B / aircraft pattern checks: rank frequently seen aircraft/identifiers using cached ADS-B heard aircraft and local event logs.
  • 24h SITREP assembly: build a last-24h SITREP with timeline highlights, key concerns, and recommended actions.
  • Anomaly scan: detect spikes, repeats/loops, template concentration, hot terms, and concern clusters.
  • Packet drafting: generate XTOC COM packet drafts for any schema and flag unknown fields as TODO.
  • Off-grid AI relay: encode/decode AIASK/AIREPLY packets (X1.50 / X1.51) so answers can move over constrained transports.

XCORE AI

How it works

Simple pipeline. Local execution.
Local XTOC/XCOM data (logs + snapshots)
  -> context extraction (AO, roster, packets, terms, ADS-B)
  -> mode instruction (Fusion / Directive / Scan)
  -> local CPU model inference (GGUF via GPT4All)
  -> operator-ready response + packet drafts
Tune model size from fast CPU baseline to larger options. Keep data roots focused for the highest signal.

Modes

  • XCORE™ Fusion: balanced operational assessment. mode: XCORE-Fusion
  • XCORE™ Directive: options + recommendation + immediate actions. mode: XCORE-Directive
  • XCORE™ Scan: rapid anomaly-oriented triage. mode: XCORE-Scan

Local model runtime and licenses

  • Local-only inference: no remote model API calls in normal operation.
  • Default model (quality-first): mistral-7b-instruct-v0.1.Q4_0.gguf.
  • Model Library: includes standard and larger options.
  • License visibility: model picker shows license name and docs include a reference table.
Operators can import local GGUF files fully offline. Optional online model download is manual and can be disabled in Settings.

XTOC and XCOM integration

XCORE™ runs a local HTTP bridge and OTA packet endpoints so XTOC/XCOM workflows can query and relay AI output.
  • Bridge endpoint: POST /v1/run for immediate analysis.
  • XTOC snapshot sync: POST /v1/xtoc/sync can keep XCORE™ grounded (roster + recent packets + cached ADS-B heard aircraft).
  • Export-for-XCORE™ (file handoff): portable snapshot when XCORE™ runs on a different machine.
  • OTA endpoints: queued asks, job polling, packet encode/decode helpers.
  • Packet templates: AIASK (X1.50) and AIREPLY (X1.51) with chunking for constrained transports.
Security: keep the bridge on 127.0.0.1 whenever possible. If you bind to LAN, restrict inbound access (trusted subnets only) and never expose the port to the public internet.

Local-first by design

  • No cloud inference requirement: runs with local GGUF models.
  • No account dependency: config, threads, and run logs remain local.
  • Operator-controlled downloads: online model download toggle is explicit and optional.

Safety and compliance notes

XCORE™ is an assistive analyst. Operators must verify outputs before transmitting or acting.
Always validate tactical conclusions and packet fields against ground truth and command protocol.

FAQ

Does XCORE™ require internet?
No for inference if a local model is installed. Internet is only needed if you choose to download models from online sources.

Does XCORE™ require Ollama?
No. This build uses local GGUF models through GPT4All and does not require an Ollama server.

Can it run on CPU?
Yes. The default model is selected for CPU usability.

How does it integrate with XTOC?
XTOC calls local bridge endpoints and displays responses with XCORE™ status language.

How does it integrate with XCOM?
Point XCORE™ data roots at local XCOM logs and use OTA packet relay for cross-team ask/reply workflows.

Get XCORE™

Purchase and download: XCORE™ – Tactical AI Analyst + XTOC Integration (Windows) – MKME Store
Downloads page (post-purchase): https://mkme.org/xcore
Community and support: https://mkme.org/discord
Pairs with: XTOC and XCOM