gameplay-recording-replay

Build a deterministic gameplay record/replay test artefact for Unity, Unreal, or Godot - record a player session, save it to disk, replay it bit-for-bit, and assert that the resulting game state match

SKILL.md

GitHub

gameplay-recording-replay

Overview

A deterministic replay artefact is a recorded session of player input or replicated state that can be played back to recreate the original game state. Three reasons games invest in replays:

  1. Regression tests. Record a scripted playthrough; replay it in CI; assert the final state matches the original. Catches physics drift, AI determinism breaks, save-format regressions.
  2. Bug repros. Players record a session; QA replays it inside the build with the bug; cert teams attach the replay to the ticket.
  3. Esports / spectator / clip sharing. Match replays surfaced in-product.

This is a workflow that produces, per engine, a working record/replay setup plus a CI regression test that runs it. The full per-engine record, replay, and CI code lives in references/engine-record-replay-apis.md; this file is the cross-engine decision surface.

The three engine surfaces differ:

Engine Recording surface Stability of public docs
Unity Input System's InputEventTrace (input-level) Public - com.unity.inputsystem@1.8 API
Unreal Replay System (replication-stream level) - DemoRec / DemoPlay / DemoStop Public - Replays in Unreal Engine
Godot No first-party replay - community pattern: deterministic RNG seed + recorded InputEvent script Community - Godot documentation does not ship a replay subsystem [author opinion, per docs.godotengine.org]

Composes with:

  • unity-test-framework, unreal-automation-system, godot-gut-tests - the host test framework the replay runs inside.
  • multiplayer-state-machine-coverage - Unreal's Replay system shares NetworkReplayStreamer with its multiplayer stack; multiplayer fixtures can replay captured sessions.
  • platform-cert-overview-reference - replays are evidence artefacts for cert (Xbox XR-003 title quality, attaching a repro to a CFR).

When to use

  • Producing a deterministic regression test for a story playthrough / level traversal / damage script.
  • Building an in-product replay feature (spectator mode, instant replay, clip sharing).
  • Producing reproducible repro artefacts for cert teams.
  • Investigating a determinism regression - "this build diverges from the previous build's replay at frame 4823."

How to use

  1. Lock determinism first for the capture level you need - fix the Step 1 sources of non-determinism (fixed timestep, seeded RNG, synchronous loads) before recording anything.
  2. Pick the capture level from the Step 2 per-engine contract table - Unity and Godot capture input (which needs determinism); Unreal captures the replication stream (which does not).
  3. Record and replay with the engine's API - full working code in references/engine-record-replay-apis.md.
  4. Choose the assertion model (final-state hash, per-checkpoint, or per-frame) and wire the replay into a CI regression test (Step 3).
  5. Embed a build hash + replay-format version in the replay header so CI can tell a real regression from a stale replay (Step 4).

Inputs

Gather before walking the workflow:

Input Where Why
Engine + version Project Determines which API surface applies
Determinism baseline Game design - is fixed-tick physics on? RNG seeded? AI deterministic? Replays only work when the game's per-frame outputs are functions of (state + input)
Recording scope Input-only? Full replication stream? Trades off file size against determinism guarantees
Target storage location Application.persistentDataPath, %LOCALAPPDATA%/<Project>/Saved/Demos, etc. Where replay files land
Replay length budget Seconds / minutes InputEventTrace buffer sizing
Assertion model Final-state hash? Per-frame? Checkpoint-only? Drives the test harness shape

Workflow

Step 1 - Establish determinism

Replays are worthless if the game is non-deterministic. Before any recording work, lock down:

Source of non-determinism Lock-down
Variable timestep / Time.deltaTime Run physics on FixedUpdate (Unity) / TickGroup (Unreal) with a fixed delta
Unseeded Random Seed every RNG at session start; record the seed in the replay header
Async load timing Force synchronous load during replay (SceneManager.LoadScene sync; AssetRegistry::Tick to drain)
Multithreaded game logic Pin to one thread or commit to ordered consumption of results
Frame-rate-dependent FX Tag visual-only systems and skip them in headless replay runs

If you can't lock determinism, replay can still be useful as a visual debug artefact, but it won't drive regression assertions.

Step 2 - Record and replay per engine

Each engine captures at a different level, with a different determinism contract. Pick the row that matches your engine, then lift the working record + replay code from references/engine-record-replay-apis.md.

Engine Capture level Default storage Determinism contract
Unity Input events (InputEventTrace, buffer-sized) Application.persistentDataPath / repo test/ Needs Step 1 - replay reproduces input; state matches only if the sim is a pure function of state + input
Unreal Replication stream (DemoNetDriver + NetworkReplayStreamer) %LOCALAPPDATA%/<Project>/Saved/Demos Reconstructs the server's streamed view - does not require local determinism
Godot Input events + RNG seed (community pattern) FileAccess path of choice Same contract as Unity - needs Step 1

The key split: input-level replay (Unity, Godot) depends on determinism; replication-stream replay (Unreal) does not, because the Unreal simulation already ran on the server and the replay just re-streams its output. This is why Unreal replays handle full multiplayer matches while Unity/Godot replays are for pinned single-player.

Step 3 - Wire replays as CI regression assertions

Per engine, the CI loop is:

  1. Check the replay file into the repo under test/replays/.
  2. CI job loads the build, plays the replay, hashes the final game state.
  3. Compare the hash against the baseline (also in repo).
  4. Mismatch → CI fails the build.

The per-engine harness code (Unity [UnityTest] PlayMode fixture, Unreal LatentIt latent command, Godot GUT fixture) is in references/engine-record-replay-apis.md.

Step 4 - Define a fail-safe for replay drift

Real-world replays drift when:

  • The game build version changed and the replay was recorded against an earlier version.
  • A non-deterministic source slipped in (new async loader, new particle system using Random).
  • The save format changed.

The harness must distinguish bug-in-build from replay-stale. Encode the build hash + replay format version in the replay header; CI rejects replays whose replay_format < current_format, then either re-records (acceptable for visual replays) or fails the build (regression test replays).

Worked example - Unity speedrun regression

Inputs:

  • Engine: Unity 6.0, Input System v1.8.
  • Determinism: Physics at FixedUpdate, 60 Hz; RNG seeded from replay header; AI scripted via deterministic state machines.
  • Recording scope: Input-only via InputEventTrace.
  • Storage: Assets/Tests/Replays/level1.trace (committed to the repo).
  • Length budget: 90 s - fits comfortably in 4 MB buffer (recordFrameMarkers = true per the API page).
  • Assertion: SHA-256 hash of GameStateRoot.Serialize() at T=90 s.

Step 1 - physics on FixedUpdate, RNG seeded from replay header, async loads forced synchronous in headless mode.

Step 2 - recorder MonoBehaviour creates InputEventTrace, enables on level start, writes to level1.trace on first level completion.

Step 3 - CI fixture (UTF [UnityTest]) reads the trace, replays it, hashes the final GameStateRoot, compares to the baseline hash committed in Assets/Tests/Baselines/level1.hash.

Step 4 - replay header includes replay_format = 2 and build_hash = <git-sha>. CI fails the build with a clear "replay stale" message when format < 2 rather than silently passing.

Anti-patterns

Anti-pattern Why it fails Fix
Recording input without seeding RNG Replay diverges by frame ~30 Step 1 - every RNG seeded; seed in replay header
InputEventTrace without Dispose() Memory leak per API page - "must be disposed of … or they will leak memory on the unmanaged (C++) memory heap" Always using / Dispose()
InputEventTrace without recordFrameMarkers = true Replay reissues events back-to-back instead of respecting timing per API page Set recordFrameMarkers = true before Enable()
Comparing per-frame state instead of checkpoints False-positive failures from float-precision drift Hash at coarse-grained checkpoints (every N seconds or per level)
Treating Unreal replays as deterministic input replays They're replication-stream replays - different determinism contract per Replays page Document the contract; don't assume input-level determinism
Storing replay format version implicitly Replay-stale failures get misdiagnosed as bugs Embed replay_format + build_hash in the header
Replays committed without their baselines Asserting nothing Commit replay.trace + baseline.hash as a pair
Replays > 30 min in CI Replay run time dominates total job time Split into checkpointed mini-replays
Replay scrubbing in test fixtures Many engines don't expose deterministic scrubbing for tests Linear playback only in regression fixtures; scrubbing is a feature, not a test surface
Replays storing PII Saved player names / chat messages in attached repros Strip PII from replay headers before sharing

Limitations

  • Input-level replay (Unity) and replication-stream replay (Unreal) are different contracts. Unity replays depend on determinism; Unreal replays reconstruct the server's view and do not. Don't conflate them in a multi-engine team.
  • Godot has no first-party replay system per the Godot documentation home; the community pattern in Step 2 is sufficient for deterministic single-player but doesn't generalise to multiplayer.
  • Unreal replays bloat with high-bandwidth replication. The .replay file under %LOCALAPPDATA%/<Project>/Saved/Demos grows with replicated-property volume - large open-world games produce GB-scale replays. Use checkpoint-only replays for long sessions.
  • Build-version coupling. Replays recorded against build N rarely play correctly on build N+1 if the netcode / save format changed. Treat the replay file as a build-pinned artefact.
  • No public common exit-code definition for replay-failed states in either Unity or Unreal - parse the test framework's result file (UTF NUnit XML or Unreal automation JSON) for the failure detail, don't rely on engine exit code.
  • No assertion API on Unreal DemoNetDriver replay completion in public docs - the spec / latent-command harness must poll the demo time and time-out. See Replays in Unreal Engine.
  • PII in replays. Player chat, gamertags (per Microsoft XR-046 modern-gamertag display rules), and inventory data can land inside replay attachments - strip before sharing with cert / publisher.

References