Peekling Research guide · v0.1

For researchers

Study a small digital presence with visible rules.

Use Peekling's inspectable Plan, Pack, and runtime boundaries as a reproducible base for interaction research.

01 · AVAILABLE TODAY

Use an inspectable runtime as study material.

Peekling provides a versioned browser runtime, immutable Plans, data-only character Packs, and explicit visitor controls. A study can self-host exact artifacts and describe each supported trigger and response.

Peekling supplies the character runtime under study. The study host supplies participant recruitment, condition assignment, surveys, datasets, and analysis.

02 · INSPECTABLE MODEL

Separate the character from the experiment host.

Study hostCondition, consent, and instrumentation
EventA bounded browser or application fact
Immutable PlanDeclared Rules and channel ownership
Visible responseValidated State, motion, or surface

The host owns study logic and measurement. The Pack owns artwork and capabilities. The Plan owns the character's response policy. Keeping those roles separate makes the experimental condition easier to describe.

03 · RECORD THE CONDITION

Archive more than a screenshot.

  1. Pin the exact runtime and Pack versions.
  2. Save the complete Plan JSON and Configuration used by the host.
  3. Record the manifest URL, atlas density, and declared atlas hash.
  4. Record browser version, viewport, device pixel ratio, and render scale.
  5. Document reduced-motion, visibility, and dismissal conditions.
  6. Self-host the same approved bytes when delivery must stay controlled.

A local working tree can differ from a published package. Treat the pinned artifact and its recorded hash as the study input.

04 · CHOOSE VARIABLES

Change one declared property at a time.

A study may compare a stable baseline with a declared change in Pack, State, timing, movement, scale, position, Event mapping, or surface copy. Keep unrelated product behavior fixed and preserve the same visitor controls in every condition.

Questions about attention, mood, learning, trust, engagement, and conversion belong in the study protocol. The engine provides observable behavior and bounded variables for testing them.

05 · PARTICIPANT CONTROL

Keep refusal and recovery inside the protocol.

  • Explain what the host records and why before collecting it.
  • Keep the shared rest menu and keyboard path available.
  • Respect reduced motion and provide a stable non-animated condition.
  • Keep task controls clear and make dismissal independent of participation.
  • Include a normal "Show again" action when the study requires recovery.

06 · MEASUREMENT BOUNDARY

Instrumentation belongs to the study host.

The Peekling runtime is telemetry-free. Diagnostics remain local and omit Event payloads, content, DOM text, selectors, and resource URLs from engine-generated metadata.

The study host defines its measurement layer, consent, data minimization, retention, security, and ethics review. Participant authorization governs every item the host collects.

07 · NEXT STEP

Define the recorded condition

Read the versioned execution contracts.

Open the specifications →