Welcome
Welcome to your Unified Cloud and AI Gateway for managing, securing, and scaling Cloud and AI operations across multiple vendors and on-prem infrastructure — from storage buckets and AI model endpoints to chat assistants and organization-wide governance.
Clouds are fragmented. AI is out of control. Flashback puts your operations back in your hands.
Flashback connects hyperscalers and all type of cloud environments through a single, consistent control layer. What sets us apart is our ability to use your own AWS, GCP (and any compatible solution, even on-prem) accounts and credentials in a Bring-Your-Own-Key (BYOK) model. You remain in full control of your cloud accounts, access keys, and encryption keys, while Flashback focuses on policy, routing, and governance on top.
With Flashback, you can connect cloud storage and AI LLMs into unified Repositories, power controlled AI assistants through a Chat Engine tailored to your teams, and manage everything through an organization-wide control plane for observability, policy, privacy, and cost control.
Whether you’re a technical founder, system architect, DevOps engineer, or data scientist, this documentation will help you onboard quickly and take full advantage of Flashback’s federated cloud capabilities.
Why Flashback?
Built for Every Stage of Growth
From first experiments in a garage to global-scale deployments, Flashback is designed to grow with you. It empowers builders, startups, and enterprises to use the same unified cloud fabric from day one — removing the friction that usually comes with scaling or switching providers.
You can start small, connecting a few storage buckets or compute instances to test your ideas, and expand seamlessly into multi-region, multi-cloud, or even decentralized deployments as your product matures. Flashback’s architecture adapts dynamically: resources can be added, migrated, or rebalanced without breaking compatibility or rewriting code.
Beyond tooling, Flashback also acts as a guide and educator.
Our technology is helping teams understand how to make cost-efficient, sustainable, and secure infrastructure choices at every stage of their journey. For new builders, it provides simple defaults and safe onboarding; for advanced teams, it offers observability, automation, and performance insights across every connected environment.
By unifying the lifecycle from experimentation to optimization, Flashback ensures that your infrastructure never becomes a constraint to your ambition — it becomes a partner in growth, evolving as fast as your vision does.
Bring Your Own Key, Keep Your Control
Flashback is designed around Bring-Your-Own-Key principles:
You connect your own cloud accounts and storage backends; Flashback does not take over or resell capacity on your behalf.
Access is granted through scoped credentials and roles that you configure and can revoke at any time.
Encryption keys, IAM roles, and provider-specific policies remain under your governance; Flashback works with them instead of replacing them.
Repository-level API keys in Flashback are layered on top of your existing providers, enforcing least-privilege access for applications, teams, and services.
This way, you gain a unified control plane across clouds without giving up ownership of your underlying infrastructure or secrets.
Unified Cloud Fabric
We are developing Flashback as a truly unified fabric that integrates storage, compute, networking, and AI services from AWS, GCP, Azure, and any hyperscaler-compatible provider through a single, consistent API. This architecture allows organizations to blend the reliability of established infrastructures with the agility of emerging ones — all within one seamless operational layer.
Behind the scenes, Flashback standardizes authentication, routing, and permissions across clouds, giving developers and operators a coherent environment to deploy, monitor, and scale workloads anywhere. With fine-grained access control and unified policy enforcement, teams can manage data placement, lifecycle, permissions, and model usage with precision — meeting the highest standards of governance and compliance.
Our platform also empowers you to operate with cost controllability in mind: resources can be dynamically routed or rebalanced based on usage, region, or provider economics, ensuring predictable and optimized spending across your entire estate. Compliance and audit features are built-in, not bolted-on, providing transparency and traceability from the first API call to enterprise-scale orchestration.
Flashback transforms multi-cloud diversity into a strategic advantage — offering a federated, policy-aware infrastructure where governance, efficiency, and freedom coexist by design.
Go Beyond in the Clouds
Flashback is the only platform that truly bridges the gap between traditional hyperscalers and the new generation of decentralized and independent clouds. Through our federation layer, you can diversify your workloads beyond AWS, GCP, or Azure — exploring innovative networks such as Akave, Nomi Cloud, Walrus, or StorJ — all within the same control plane.
For developers, Flashback offers a seamless way to build, test, and deploy across both centralized and decentralized backends without changing tools or APIs.
For emerging cloud providers, it provides global visibility, standardized integration, and fair access to enterprise demand.
By connecting these two worlds, Flashback enables a more open, resilient, and cost-efficient cloud ecosystem — where innovation and interoperability thrive side by side.
What You Will Find Today
Flashback already provides the foundation of your federated cloud environment, giving you unified access to the essential building blocks of multi-cloud and multi-AI operations.
Buckets are remote object-storage endpoints such as an AWS S3 bucket, Google Cloud Storage bucket, Microsoft Azure Blob container, or any S3/GCS-compatible backend—that you register with Flashback. Each bucket is a representation of that vendor’s storage in your Flashback account: it keeps its native configuration and permissions while gaining unified access control, routing, and policy enforcement through the Flashback platform.
Flashback lets you connect AI LLM providers (such as OpenAI-compatible endpoints) into Repositories, so your applications can call multiple models through a unified, policy-aware interface. You keep using your own AI provider accounts and keys, while Flashback standardizes routing, access modes, and usage boundaries for different projects, teams, and environments.
On top of these connected AI LLMs, the Flashback Chat Engine provides a controlled interface for building AI assistants that can be scoped to specific departments, groups, or use cases inside your organization. You can define which Repositories, models, and data each assistant can access, apply guardrails and policies per team, and give users a clean chat UI that is backed by enterprise-grade governance rather than ad hoc scripts.
Flashback includes an organization management interface where you can oversee how resources, Repositories, and chat assistants are used — from very small teams to large enterprises. It is designed to support observability (who uses what, where, and when), policy and permission management across groups, and governance features aimed at better privacy and cost control. This gives platform owners, security teams, and finance stakeholders a shared view of AI and cloud usage across the organization.
Repositories are workspace-level containers that group together multiple resources under a single API interface. You can attach several types of resources — today, this includes storage buckets (S3, GCS, Azure Blob, S3-compatible) and AI LLM endpoints (OpenAI-compatible) — and expose them through a chosen endpoint type (S3/GCS/Azure Blob for storage or OpenAI for AI LLMs). Through that unified interface, you generate Repo API keys, manage credentials, and define fine-grained permission scopes (READ, WRITE, READ/WRITE, ADMIN) that apply to all compatible resources attached to the Repo. These repository API keys sit on top of your own provider credentials in a BYOK model: Flashback enforces access policies, while you keep ownership and control of the underlying accounts, keys, and model subscriptions. Over time, Repositories will also become the place where you attach additional resource types (machine learning endpoints, SQL, VMs, and more) and manage keys and policies around that bundle.
Flashback-managed endpoints that connect your data and workloads across providers. They translate native storage or compute protocols into Flashback’s APIs, collect quality-of-service metrics, and enforce SLAs. Bridge Nodes are also the foundation of our decentralized infrastructure layer — contributing to reputation, staking, and automated optimization.
What’s Coming Next
The Flashback platform is evolving into a complete federated cloud operating system. Beyond unified object storage, AI LLM connectivity, and chat/organization management, our roadmap expands into:
AI Engine Aggregation: Integrate, route, and orchestrate inference workloads across multiple AI providers under one control layer.
Agentic Self-Management: Let intelligent agents observe performance, forecast demand, and autonomously rebalance workloads for optimal cost and uptime.
Multi-Cloud Compute & Serverless Functions: Deploy VMs, containers, and functions across clouds and DePIN networks as easily as you manage storage today.
Structured and Specialized Data Layers: Extend your repositories to support databases, analytics engines, and memory-optimized compute surfaces.
DePIN-Native Governance & Incentives: Participate in on-chain validation, reputation scoring, and staking-based rewards that align performance with transparency.
Flashback is not just a multi-cloud interface — it’s a federated foundation for the next generation of computing, where AI, automation, and decentralized infrastructure converge to give every builder, researcher, and enterprise the same superpowers once reserved for hyperscalers
Getting Started
Create a Bucket Register your first storage endpoint under Storage > Buckets. Refer to Create a Bucket for details.
Create a Repository Aggregate one or more resources (buckets and/or AI LLMs) into a repository under Storage > Repositories. See Create a Repository and issue READ/WRITE keys in your repository’s API Keys tab to start interacting programmatically.
Integrate with Your Tools Point your S3/GCS/Azure-compatible client or OpenAI-compatible client at Flashback’s Bridge endpoint and begin uploading/downloading data or calling models through your Repositories.
Connect Your AI LLMs (BYOK) Under the AI/LLM configuration section, plug in your own provider credentials (e.g., OpenAI-compatible keys, Bedrock/Vertex/Azure keys where applicable) and attach those models to a Repository. You keep full control over your AI provider accounts and keys; Flashback only standardizes access, routing, and permissions on top.
Explore the Chat Engine Open the Chat or Assistants area, create a test assistant bound to one or more Repositories, and start chatting. Use this to validate how your connected buckets and AI LLMs behave behind a controlled interface (per-team context, permissions, and guardrails).
Start Managing Your Organization Go to the Organization / Settings section to invite team members, create groups (departments, squads, projects), and assign access to Repositories and assistants. From there, you can explore observability, basic policies, and governance options that help you keep privacy, security, and costs under control as usage grows.
Our Community
Join Our Community: Have questions or suggestions? Contact [email protected] or chat with us:
Next Steps
Navigate the left menu to find detailed instructions:
Last updated
Was this helpful?