Spiritual Resonance Engine

Make space for resonance. Leave room for meaning.

Comment

Inspiration

Spiritual Resonance Engine creates a gentle, safety-first way to explore meaning in a moment without turning people into data profiles. It maps a small amount of abstracted context to a Scriptural theme, a curated passage reference, an invitational reflection, and an optional micro-action. Boundaries were informed by the project’s responsible-capture brief and by public discussion about legible sensing, including Meta's AI-glasses norms discussion and a related Meta-regulation reflection. [https://newsletter.aicollective.com/p/meta-s-cto-says-norms-will-sort-out-ai-glasses-norms-need-a-witness] [https://lindsaygross1.substack.com/p/in-2021-metas-regulator-asked-if]

What it does

Accepts only declared, abstracted moment metadata. Applies a safety gate before theme mapping and validates the final response again. Returns a bounded, structured response: theme, Scripture reference, pastoral reflection, micro-action, and safety information. Provides strict, expanded, and developer safety profiles while prohibiting person-level inference across all of them. Includes a clean MapLibre interface. City/US ZIP search centers the map only; map queries and coordinates are never sent to the engine or stored.

How we built it

The engine is a deterministic Python pipeline with small, auditable modules for safety filters, theme mapping, curated Scripture selection, reflection prompts, and gentle nudges. A FastAPI wrapper exposes the engine through POST /resonate; a responsive MapLibre UI makes the experience easy to demonstrate.

OpenAI Codex with GPT-5.6 Terra accelerated the implementation of the schema boundary, safety and content modules, API, MapLibre UI, and test suite. GPT-5.6 and Codex were used for development; the live engine makes no model or OpenAI API calls at runtime.

Responsible-by-design

The central rule is: describe the moment, never the person. The engine rejects raw sensing data, precise location traces, identity, demographic, health, political, relationship, diagnostic, predictive, and moral-scoring inputs. In sensitive contexts, it uses curated references and gentle templates designed to avoid coercive, shaming, or evaluative language.

Try it

The live demo is available at https://spiritual-resonance-engine.onrender.com/. The project is open source and includes setup, tests, deployment instructions, and a public input schema in the repository.

Challenges we ran into

  • Defining a useful experience without collecting sensitive or person-level data.
  • Designing safeguards that keep the engine focused on an abstracted moment—not an identity, diagnosis, score, or prediction.
  • Ensuring reflections and micro-actions remain invitational, gentle, and free of judgmental, harsh, or coercive language.
  • Separating the map’s visual context from the engine so coordinates are never stored or sent for interpretation.

Accomplishments that we're proud of

  • Developing clear safety profiles—strict, expanded, and developer—that adjust permitted context while preserving the same prohibition on person-level inference.
  • Turning responsible-inference principles into enforceable schema validation, input/output safety filters, curated content rules, and transparent rationale.
  • Building a complete, open-source experience: deterministic Python engine, FastAPI service, privacy-preserving map helper, and polished MapLibre interface.

What we learned

  • Deterministic rules remain essential for responsible inference and capture. Clear boundaries are not a limitation; they make behavior more auditable, predictable, and trustworthy.
  • Privacy needs to shape product architecture from the beginning, especially when a visual interface could otherwise imply location tracking or hidden data collection.
  • Safety is strongest when it is visible in the product experience, not only described in documentation.

What's next for Spiritual Resonance Engine

  • Continue refining safety profiles, capture rules, and transparent rationale for every response.
  • Deepen the core engine’s curated themes, Scripture selection, reflections, and gentle micro-actions while keeping its deterministic safety boundary.
  • Add user-facing options that provide meaningful control without introducing profiling or person-level inference.
  • Support multiple client types through the API, while preserving the same schema and safeguards.
  • Explore a larger application powered by the Spiritual Resonance Engine, with this safety-first engine remaining its foundational layer.

Built With

Share this project:

Updates