Skip to documentation
Screebits Docs
Flow Editor
Open Screebits
  • Screebits
  • Flow Editor documentation
  • Blueprints
  • Your first steps in Screebits
  • Workspace, strategies and account
  • Get started with the Flow Editor
  • Build an RSI Momentum Alert
  • Flow Editor concepts
  • How a strategy starts and continues
  • Control and data connections
  • Nodes, parameters, outcomes, and ports
  • Plans and feature availability
  • Saves, versions, and checkpoints
  • Tactical and Portfolio strategies
  • Validation and readiness
  • Flow Editor mode guides
  • Build Portfolio strategies
  • Build Tactical strategies
  • Flow Editor guides
  • Add a Quant Hub metric
  • Configure a node
  • Configure Settings and visual preferences
  • Connect nodes
  • Create a strategy
  • Create checkpoints and restore versions
  • Reset a Flow safely
  • Debug and replay a Backtest
  • Fix validation issues
  • Use keyboard shortcuts
  • Manage variables
  • Run a Backtest
  • Save and resolve conflicts
  • Work on tablet and mobile
  • Use Quick add
  • Use the Inspector
  • Add, move, duplicate, and remove nodes
  • Flow Editor examples
  • Portfolio rebalance with risk limits
  • Tactical order and position monitoring
  • Use Backtest Center
  • Read a Backtest result
  • Compare and export Backtests
  • Explore Quant Hub
  • Metric parameters, charts and overlays
  • Flow Editor node reference
  • Flow Editor troubleshooting
  • Troubleshoot activation and validation
  • Troubleshoot Backtests and replay
  • Troubleshoot compact layouts
  • Troubleshoot connections and data types
  • Troubleshoot plans and locked features
  • Troubleshoot Quant Hub in the Flow Editor
  • Troubleshoot save, offline, and conflicts
  • Backtests and responsible use
  • Flow Editor glossary
  1. Screebits
  2. /Flow Editor documentation
  3. /Flow Editor concepts
  4. /Saves, versions, and checkpoints
IntermediateLast verified Sep 30, 2026

Saves, versions, and checkpoints

Understand dirty state, autosave, conflicts, history, and non-destructive restore.

First strategy · Step 9 of 10

On this page
  1. Save states
  2. Versions and recovery entries
  3. Restore is non-destructive

Saving, version history, and checkpoints protect different parts of your work. A save state tells you whether the current edit reached storage. A version records a snapshot in history. A named checkpoint marks a state you want to find again. Do not confuse a historical preview with the editable current strategy.

Save states

  • Dirty means the graph has local changes.
  • Saving means storage is in progress.
  • Saved means the latest accepted changes are stored.
  • Offline, error, or conflict require recovery before you assume the latest edit is stored.

Autosave runs while you edit, and manual Save is available whenever saving is allowed. Before leaving, verify the visible Saved state for your latest change. If the state is Offline, Error, or Conflict, keep the tab open and follow save recovery. A Backtest result refers to the version tested; editing the current graph afterward does not rewrite that result.

Try it here

Practice a save conflict

Understand dirty state, autosave, conflicts, history, and non-destructive restore.

Exercise steps
  1. Change the threshold on Condition, then save a local checkpoint.
  2. Choose a simulated save outcome. Recovery actions appear when that state requires them.
Event TriggerData SourceIndicatorConditionAlert
Graph preview — a diagram of the example

Versions and recovery entries

History can include the current version, recovery entries, named checkpoints, and technical entries. Ordinary saves and autosaves are recovery records, not named checkpoints. The version selector groups categories; operational saves may be collapsed, and older items use bounded pagination or search. Version retention depends on the current policy and plan, so do not assume every recovery save or checkpoint lasts forever.

Previewing an older version opens it in read-only mode. It does not replace the editable graph; editing and autosave are unavailable there. Use Return to Main to resume current work. If a Backtest opens an exact tested snapshot, treat that preview as historical evidence rather than as a place to modify the strategy.

Restore is non-destructive

Restoring creates a new current version from selected historical content. Start the restore from the Versions tab on the strategy detail page when the selected entry offers the action; the historical editor preview is for inspection. Save or resolve current local changes first. The restore confirmation may disclose retention impact that needs acknowledgment. After success, open the current editor and verify its nodes and settings. The restore operation does not rewrite an old Backtest or silently turn its snapshot into a new run.

For example, save a checkpoint named RSI 55 baseline, change the threshold to 60, and save. The current strategy is now 60, while the named checkpoint lets you inspect the earlier 55 configuration. A restore would create a new current version based on 55; it would not erase the fact that you later tried 60. The checkpoint dialog previews any applicable retention effect before creation.

Related documentation

  • Save and resolve conflictsProtect local edits, recover from offline or error states, and choose the correct version after a conflict.
  • Create checkpoints and restore versionsCreate a named recovery point, preview history safely, and restore without deleting older versions.
  • Troubleshoot save, offline, and conflictsProtect local edits and recover when the editor cannot confirm the latest saved version.
PreviousFix validation issuesNextRun a Backtest

On this page

1 / 3
  1. Save states
  2. Versions and recovery entries
  3. Restore is non-destructive

Practical guidance for strategies, historical simulations and metric research in Screebits.

Responsible useGlossary