genblaze
Genblaze is an open source Python SDK for orchestrating generative AI media pipelines across video, audio, and image providers with built in provenance for ever
What is genblaze?
Genblaze is an open source Python SDK for orchestrating generative AI media pipelines across video, audio, and image providers with built in provenance for every output.
Overview of genblaze
What it is
Genblaze is an open source Python SDK by Backblaze for building and orchestrating generative AI media pipelines across video, audio, and image providers. It provides a unified Pipeline API that spans providers like OpenAI, Google, Runway, Luma, ElevenLabs, and Stability Audio, as well as models served through platforms such as GMI Cloud and NVIDIA NIM. Every run produces a canonical provenance manifest that can be embedded directly into media files (.mp4, .png, .mp3, etc.) and persisted to Backblaze B2 or any S3-compatible store. The manifest includes a SHA-256 hash for integrity verification via Manifest.verify().
Key features
- Provenance by default: Every run yields a canonical manifest that is deterministic, embeddable into media files, or persisted alongside the asset. Outputs become SHA-256-covered when providers return bytes or
ObjectStorageSinktransfers them into durable storage. - One pipeline, many providers: Eleven adapters across video, image, audio, and chat behind a single
Pipeline/StepAPI. Providers can be swapped by changing one line. - Storage is first-class:
S3StorageBackend.for_backblaze('bucket')ships durable, credential-free asset URLs and content-addressable layouts. Works with any S3-compatible store (AWS S3, Cloudflare R2, MinIO). - Fallback chains and conformance:
fallback_models=[...]retries onMODEL_ERROR; CI-gradeprobe_modelsand provider-contract tests catch upstream drift. - Replayable runs: Every manifest captures full provenance (provider, model, prompt, params, timestamps) so a run can be reconstructed via
genblaze replay manifest.jsonor by feeding the canonical params back into a Pipeline.
Use cases
- Building generative media pipelines that require provenance tracking for every output.
- Orchestrating workflows across multiple AI providers (video, image, audio) without rewriting orchestration code.
- Persisting generated assets and their provenance manifests to cloud storage (Backblaze B2 or S3-compatible) with durable, credential-free URLs.
- Implementing fallback chains to handle provider errors and ensure pipeline reliability.
- Replaying or auditing past runs using the captured manifest data.
Getting started
Install the core package and a provider adapter: bash pip install genblaze-core genblaze-gmicloud genblaze-s3 Set environment variables for API keys and B2 credentials: bash export GMI_API_KEY='gmi-...' export B2_KEY_ID='...' export B2_APP_KEY='...' Run a quickstart example that generates a video, persists it and its provenance manifest to Backblaze B2, and verifies the hash: python from genblaze_core import Modality, ObjectStorageSink, KeyStrategy, Pipeline from genblaze_gmicloud import GMICloudVideoProvider from genblaze_s3 import S3StorageBackend storage = ObjectStorageSink( S3StorageBackend.for_backblaze('my-bucket'), key_strategy=KeyStrategy.HIERARCHICAL, ) result = ( Pipeline('my-first-pipeline') .step( GMICloudVideoProvider(), model='seedance-2-0-260128', prompt='A drone shot soaring over a coastal city at golden hour', modality=Modality.VIDEO, duration=10, aspect_ratio='16:9', ) .run(sink=storage, timeout=600) ) print(f'Asset URL: {result.run.steps[0].assets[0].url}') print(f'SHA-256: {result.run.steps[0].assets[0].sha256}') print(f'Manifest: {result.manifest.manifest_uri}') print(f'Hash: {result.manifest.canonical_hash}') print(f'Verified: {result.manifest.verify()}') For more installation options and provider adapters, see the repository README.
License and pricing
The project is licensed under the MIT License and is open source. No pricing information is provided beyond the open source nature.
Based on the project's public README and repository metadata.