Practical overview
When it fits
Use Genblaze when an application needs a Python orchestration layer for generative media providers and a manifest recording how pipeline outputs were produced.
Who it is for
- Python developers building applications that combine generative video, image or audio providers.
- Teams that need workflow-level records of provider, model, prompt, parameters and timestamps for generated media.
Documented uses
- Compose multi-step workflows that generate video, images or audio across provider adapters.
- Store generated media and its provenance manifest using Backblaze B2 or an S3-compatible store.
- Retain pipeline parameters and output hashes for review or replay.
Requirements & limits
The project is open source under the MIT license. Its documented requirement is Python 3.11 or later; the B2-backed example also requires provider and storage credentials.
- Provider adapters are opt-in packages, so select and install the integrations required by the workflow.
- The documented storage package supports Backblaze B2 and S3-compatible storage.
- Manifest verification checks the manifest hash and requires a valid SHA-256 declaration for each output asset; URL-only outputs need their bytes fetched and hashed separately.
- The project documentation says manifest hashing alone is tamper-evident in trusted storage and recommends a signer or C2PA when adversarial verification matters.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Set up a B2-backed GMI Cloud video pipeline
Run the documented video-generation example and store its output and provenance manifest in Backblaze B2.
Before you start
- Python 3.11 or later
- Packages: genblaze-core, genblaze-gmicloud and genblaze-s3
- GMI_API_KEY, B2_KEY_ID and B2_APP_KEY credentials
- A Backblaze B2 bucket
Steps
Install the three documented packages with pip.
Set the GMI_API_KEY, B2_KEY_ID and B2_APP_KEY environment variables.
Create an ObjectStorageSink using S3StorageBackend.for_backblaze and a bucket name.
Build a Pipeline with GMICloudVideoProvider, a prompt and video parameters, then call run with the storage sink.
Read the generated asset URL, SHA-256, manifest URI and canonical hash from the result; call manifest.verify() as shown.
Expected outcome
The example returns a generated video asset stored through the B2-backed sink and a provenance manifest uploaded alongside it.
Guide limitations
- This recipe uses the GMI Cloud provider and requires its API key.
- The shown verification checks the manifest hash and declared asset hash; URL-only asset bytes require separate fetching and hashing.
- The guide is based on project documentation, not an execution test.
Build and verify a local manifest
Follow the documented local quickstart to create and verify a manifest without external calls.
Before you start
- Python 3.11 or later
- The project's examples/quickstart_local.py example
Steps
Open the documented examples/quickstart_local.py example.
Use that example to build a manifest locally.
Follow the example's manifest verification flow.
Expected outcome
A locally built manifest is verified without making external calls.
Guide limitations
- The local example does not describe generating media through an external provider.
- The guide is based on project documentation, not an execution test.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 14 October 2026
- Original website title
- backblaze-labs/genblaze · Official homepage title · retained in its source language