GUIDES, WORKING EXAMPLES & DOWNLOADS

Learn by
building.

Follow the standard from a recorded memory to a store that can cite — or honestly abstain. Feeling stays on SensAffect, including the RealSeries runtimes already built (Detection, Triggers, Actions). Neither path requires the other. Together they unlock the full atom.

Open the live lab ↗ RealSeries ↗ Library ↗ Operate ↗ SensAffect’s matching page ↗

1 / BUILD YOUR FIRST MEMORY ATOM

Build your first
Memory Atom.

Start with a memory record in the store. Feeling may stay empty. Add a sourced person, place or situation if you know it. Keep context separate from any interpretation — and do not ask MMRY to invent a feeling.

  1. 01
    Read the core definitionSame chapter as SensAffect step 1. Three parts. MMRY is competent for the memory side.
  2. 02
    Use the logical schemaSame schema as SensAffect step 2. Identify which fields the store persists.
  3. 03
    Open the cite or abstain labMMRY twin of the proximity lab. Try a content hit versus a feeling word versus denied authority.
  4. 04
    Download the lab JSONNotice hashes and empty or separate feeling. Influence and VAD live on the SensAffect side of the same atom.

2 / EXPLORE CONTEXTUAL SIMILARITY

Explore contextual
similarity.

Download the two-memory fixture. The records can share a kind of place in the world and still be two memories. On SensAffect, matching context does not make emotions identical. Here, matching context does not merge records, and feeling text is not an index.

  1. 01
    Compare context entity labels, not names aloneSame rule as SensAffect. Kitchen is not Garden. Two IDs stay two memories.
  2. 02
    Observe that feeling is not searchedSensAffect asks you to observe VAD differences. MMRY asks you to ask “calm” and watch an abstain.
  3. 03
    List shared context as explanation, not a scoreDo not invent a similarity percentage. Weighting and calibration remain unestablished — same warning as SensAffect.

Expected result Two independent memories. A cite from content (“kettle”, “garden”). An abstain from feeling-only words (“calm”). An abstain when authority is off, even on a content hit.

SAMPLE PROJECT

Cite or abstain lab.

A small, dependency-free browser project — the MMRY twin of SensAffect’s context proximity lab. Open it live or locally, run a query, export the turn. It demonstrates recall rules, not a production search engine and not an emotion estimator.

Open the live lab ↗ Download starter project ↓ SensAffect sample project ↗

REFERENCE DOWNLOADS

The same contract.
The store’s extras.

  1. 01
    Complete specification · MarkdownShared with SensAffect. Canonical edition 1.0.
  2. 02
    Logical schemaShared. Illustrative, not a complete runtime validator.
  3. 03
    Recall fixture · JSONMMRY twin of the context similarity fixture. Synthetic content only.
  4. 04
    Managed API · MarkdownMMRY-only: session, CSRF, remember, recall, authority.
  5. 05
    JavaScript clientMMRY-only SDK. SensAffect ships Omni-CNN for sensors.
  6. 06
    RealSeries catalogDetection, Triggers, Actions. RealFeel Ambient. This host remembers the receipt.
  7. 07
    EMC²THRPY undercurrent: cosine of feeling silhouettes. Not cite-or-abstain. Not this Express process.
  8. 08
    Documents libraryOriginal RealFlow offer. THRPY Affective Computing PDF. Mathematics markdown.
  9. 09
    RealFlow and FloAncestor lab. Named spine. RealAction tool families.
  10. 10
    Build and operateAPIs, MCPs, plugins, skills, CLIs. Live versus needed.

The schema is illustrative, not a complete runtime validator. Samples use synthetic content and demonstration values. Read the methods that still require calibration before assigning real scores. Clinical certification is out of scope.