Set up resumable uploads for Flutter
The FastPix Flutter Resumable Uploads SDK helps you efficiently upload large files from Flutter apps by splitting them into chunks and also gives you the ability to pause and resume your uploads.
How resumable uploads work through chunking
Resumable uploads can be effectively handled through a technique called chunking. This method involves breaking down large files into smaller, more manageable pieces, or “chunks.” Here’s how chunking works:
- Divide the file: Large files are split into smaller, manageable chunks (e.g., 16 MB by default).
- Upload individually: Each chunk is uploaded separately. If a chunk fails, only that specific chunk needs to be re-uploaded.
- Resume capability: If the upload is interrupted, you can resume from the last successfully uploaded chunk instead of starting over.
This approach is important because:
- It reduces the risk: Uploading smaller chunks minimizes the risk of failure. If a chunk fails to upload due to network issues, only that specific chunk needs to be re-uploaded, not the entire file.
- Improves performance: Smaller chunks can be uploaded more quickly and efficiently, especially on slower connections, as they require less time to transfer.
- Easier management: Chunking allows for better tracking of upload progress, making it easier to implement features like pause and resume.
Step 1: Install the Flutter SDK
Add the dependency to yourpubspec.yaml
Add FastPix’s Flutter library to the dependencies block of your pubspec.yaml file.
Install dependencies
Run the following command to install the dependencies:
Step 2: Create an upload URL
In order to upload a video, you will need a signed upload URL.
To get this signed URL, you’ll need a valid Access Token and Secret Key. See the Basic Authentication Guide for details on retrieving these credentials.
Once you have your credentials, use the Upload media from device API to generate a signed URL for uploading media. After fetching the the signed URL you can continue with integrating the SDK into your application.
The example below gives you an idea about how to get the signed URL using the upload media from device API. You can use the example directly or also refer to our upload videos directly guide.
Step 3: Start your upload
The Flutter SDK provides a builder pattern for easy configuration and initialization:
Builder configuration options
Step 4: Monitor upload events
The SDK provides comprehensive callback methods to monitor upload progress and handle various events:
Progress model
The progress callback provides a ProgressModel with the following properties:
uploadPercentage: Progress percentage (0.0 - 100.0)currentChunkIndex: Current chunk being uploadedtotalChunks: Total number of chunksstatus: Current upload status (e.g., “splitting_chunks”, “uploading_chunks”, “completed”)fileSize: Total file size in bytesuploadedBytes: Number of bytes uploaded so far
Step 5: Manage video uploads
You can control the upload lifecycle with the following methods:
Pause an upload
Resume an upload
Abort an upload
Check upload status
Detailed usage example
The following example gives an overview of integrating the FastPix Flutter Uploads SDK into your project, enabling you to build a fully customized upload interface:
Features
Core features
- Chunking: Files are automatically split into chunks (default chunk size is 16MB).
- Pause and resume: Allows temporarily pausing the upload and resuming after a while.
- Retry: Uploads might fail due to temporary network failures. Individual chunks are retried with exponential backoff to recover automatically from such failures.
- Lifecycle event listeners: Provides real-time feedback through various upload lifecycle events.
- Error handling: Comprehensive error management to notify users of issues during uploads.
- Customizability: Options to customize chunk size and retry attempts.
Advanced features
- Network health monitoring: Automatically detects network connectivity changes and handles offline scenarios.
- Upload lock management: Prevents multiple concurrent uploads from the same service instance.
- Progress tracking: Detailed progress reporting with chunk-level information.
- File validation: Built-in file size and format validation.
- Memory efficient: Streams file chunks without loading the entire file into memory.
BEST PRACTICES
Chunk Size: Use appropriate chunk sizes based on your target platform and network conditions. 16MB is a good default for most scenarios.
Error Handling: Always implement proper error handling to provide meaningful feedback to users.
Progress Updates: Use progress callbacks to update your UI and keep users informed about upload status.Network Monitoring: The SDK automatically handles network connectivity changes, but you may want to add additional network monitoring for better UX.
Memory Management: For very large files, consider implementing additional memory management strategies.
Retry Configuration: Adjust retry settings based on your network environment and requirements.
Troubleshooting common issues
Upload fails immediately
Check if the signed URL is valid and not expired.
Chunks fail to upload
Verify network connectivity and adjust retry settings.
Memory issues with large files
The SDK handles memory efficiently, but ensure your app has sufficient memory allocation.
Progress not updating
Make sure you’re properly implementing the progress callback.
DEBUG INFORMATION
Enable debug logging to get detailed information about upload operations.
- The SDK automatically logs debug information when running in debug mode.
- Check the console output for detailed upload information
For more information and support, contact us.
Changelog
All notable changes to this project are documented in this file.
2.0.0
Spec correctness — GCS resumable protocol compliance
- Parse the
Range:response header on 308 responses. The client now resyncs its cursor to the byte offset GCS actually committed instead of blindly advancing to the end of the chunk it sent. Fixes silent data corruption on flaky networks where the server partially commits a chunk before returning 308. - Status-query path (
Content-Range: bytes */<total>). After any transient failure / timeout / network loss / signed-URL refresh, the SDK now asks GCS for its true cursor before re-uploading. Available as_resyncCursorFromServer()internally and viarefreshSignedUrl(...)publicly. - Any 2xx is now terminal success. Previously only HTTP 200 finalized the upload; 201 / 204 fell through to the retry path.
- Empty trailing chunk guard. When the local cursor reaches EOF but no terminal 2xx has been observed, the SDK queries the server rather than PUT-ing a zero-byte (and inverted-Range) request.
Retry policy
- Real exponential backoff with jitter.
2s, 4s, 8s, 16s, 30s (cap)with ±25% jitter. Previously linear (2s, 4s, 6s…) with no jitter — no longer prone to thundering-herd on shared-backend incidents. - HTTP-status-aware retry classification. 4xx (other than 408/429) is no longer retried. 408/429/5xx are retried; everything else surfaces as a permanent failure.
- Stop swallowing timeouts and
DioExceptionType.unknown. Connection errors, send/receive timeouts, and unknown transport faults are now classified as transient and routed through the retry path. - Retry timers are tracked and cancelled on
dispose(),abortUpload(), andreset(). Stray retry callbacks no longer fire into stale state. DioExceptionType.badCertificateis treated as permanent (cert pinning / MITM situations should not be retried).
Concurrency / state
- Pause and abort no longer surface through the error stream.
Previously the SDK emitted
UploadError('Upload Paused')andUploadError('Upload Aborted')viaonError, which led consumers to treat user-initiated pause as a failure (and disable the Resume button). Pause and abort are now communicated only through the dedicatedonPause/onAbortcallbacks and the progress event with the appropriateUploadStatus. - De-singletoned
VideoUploadProgress. Was a process-wide static class whose callbacks were overwritten by every new uploader — two concurrent uploads in the same app would cross-wire their callbacks. Now per-instance. - De-singletoned
VideoUploadRetry. Per-instance retry controller owns its own pending timer. - First-network-event swallow fixed. The
_isFirstTimeflag no longer drops the first connectivity event, so an upload kicked off while offline can be auto-resumed when the network returns.
API surface (breaking changes — see “Migration” below)
uploadVideo()now returns aFuture<void>that actually resolves when the upload finalizes (or rejects withUploadErroron permanent failure / abort). Previously the future resolved immediately after the first chunk was scheduled.progressStreamanderrorStream— broadcast streams on the uploader for callers that want more than one listener or prefer streams over callbacks. The legacyonProgress/onErrorcallbacks still work.isUploading()now honors terminal failure state — returns false after a permanent failure instead of staying true forever.onUrlRefresh: Future<String> Function()— builder hook called automatically when the SDK detects an expired signed URL (HTTP 401 / 403 / 410). Mint a fresh URL and the upload resumes from the server’s committed cursor against the new URL.refreshSignedUrl(String)— manual / proactive URL replacement on the uploader..observeAppLifecycle()— opt-in builder flag. Attaches aWidgetsBindingObserverthat auto-pauses on background and resumes on foreground. Does NOT enable true background uploads (that needs platform-level integration), but leaves the resumable session in a clean state for when the user returns.- Builder default
maxRetriesreconciled with the uploader default (both now 5). Removed the dead_builderMaxRetriesfield.
Memory / performance
- Killed the double-copy in
VideoUploadChunker.readFileChunk.Uint8List.fromList(raf.read(...))is replaced with the directraf.read(...)return — saves a 16 MB copy per chunk. Uint8List.sublistViewin the progress stream in place ofsublist. For a 4 GB upload that’s ~1M fewer heap allocations.
Tests
- 32 unit tests covering chunker math, file-chunk read edge cases,
exponential-backoff math (doubling, cap, jitter bounds, never-negative),
GCS
Range:header parsing, and HTTP-status classification (200 / 201 / 204 / 308 / 308-with-Range / 400 / 403 / 408 / 429 / 500).
Migration from 1.x
- If your code relied on the static
VideoUploadProgress.emitProgress(...)/VideoUploadProgress.setupCallbacks(...)access path, switch to per-instance methods onFlutterResumableUploads(or use the newprogressStream/errorStream).
1.0.1
Documentation & Homepage URL Update
- Updated
homepageinpubspec.yamlfromhttps://www.fastpix.io/tohttps://www.fastpix.com/. - Updated documentation links in
README.md(Basic Authentication, Upload media from device) fromdocs.fastpix.iotodocs.fastpix.com. - Updated documentation link in the GitHub issue template (
.github/ISSUE_TEMPLATE/question_support.md) fromdocs.fastpix.iotodocs.fastpix.com.
1.0.0
Initial Release - Flutter Resumable Uploads SDK
A robust Flutter package for uploading large video and audio files with enterprise-grade features.
Core Features
- Chunked Upload System: Automatically splits large files into configurable chunks (default 16MB) for reliable uploads
- Resumable Uploads: Pause and resume functionality with state persistence across app sessions
- Network Resilience: Automatic retry mechanism with configurable retry attempts and delays
- Advanced Chunk-Level Retry Tracking: Individual retry tracking per chunk to prevent app sluggishness
- Real-time Progress Tracking: Detailed progress updates with chunk-level information and percentage completion
- Network Monitoring: Automatic detection of network connectivity changes with smart resume logic
- Comprehensive Error Handling: Detailed error reporting with specific error codes and messages
- Advanced Logging System: Configurable logging with multiple levels (DEBUG, INFO, WARNING, ERROR)
Architecture & Design
- Builder Pattern: Clean, fluent API for easy configuration and setup
- State Management: Robust state tracking for upload progress, network status, and retry attempts
- Modular Design: Well-organized codebase with separate modules for core, network, models, and utilities
- Memory Efficient: Proper resource management with dispose and reset capabilities
- Thread Safe: Upload lock mechanism to prevent concurrent upload conflicts
Technical Capabilities
- File Validation: Comprehensive file size and format validation
- Signed URL Support: Secure uploads using pre-authenticated URLs
- HTTP Status Handling: Proper handling of 308 (Partial Content) and 200 (Complete) responses
- Cancel Token Integration: Dio-based cancellation for clean upload termination
- Upload Statistics: Detailed logging of upload metrics and retry statistics
- Network Health Monitoring: Real-time network connectivity monitoring using connectivity_plus
Developer Experience
- Fluent API: Easy-to-use builder pattern for configuration
- Callback System: Comprehensive callback support for progress, errors, pause, and abort events
- Debug Information: Rich debugging capabilities with detailed state information
- Error Recovery: Intelligent error handling with automatic retry and manual recovery options
- Documentation: Comprehensive documentation with usage examples and best practices
Configuration Options
- Chunk Size: Configurable chunk size (default: 16MB)
- Retry Settings: Customizable retry attempts and delay intervals
- File Size Limits: Optional maximum file size validation
- Logging Control: Enable/disable logging with custom log levels and tags
- Network Timeouts: Configurable network timeout settings
Security & Reliability
- Secure Uploads: Signed URL-based authentication
- Data Integrity: Proper chunk validation and error checking
- Resource Management: Automatic cleanup and memory management
- State Persistence: Upload state tracking for reliable resume functionality
Dependencies
- dio: ^5.8.0+1 - HTTP client for network operations
- connectivity_plus: ^6.1.4 - Network connectivity monitoring
- internet_connection_checker: ^3.0.1 - Internet connection validation
Use Cases
- Large video file uploads in mobile applications
- Audio file uploads with progress tracking
- Media uploads requiring pause/resume functionality
- Applications requiring network resilience
- Enterprise applications needing detailed upload analytics