CTO Update: What Akave Built in March [2025]

Building the Backbone of Decentralized Data

A little over a year ago, Akave was just an idea—something we knew needed to exist, but hadn’t seen built right. Now, it’s real. It runs. It stores data. And people are using it.

In the last 12 months, we’ve taken Akave from zero to something real:

• We spun out of the PLVS venture studio and incorporated.
• Launched our brand and website (hi, Akave.ai).
• Raised $3.45M for our pre-seed round from Protocol Labs, Lightshift, Blockchain Builders Fund, and Blockchange.
• Dropped our Yucca testnet at Devcon in Bangkok.
• Onboarded paid customers like Intuizi and Heurist.ai.
• Hit 29k+ signups with 55+ projects building on our storage layer.
• And, we just shipped the first-ever decentralized storage integration for Snowflake.
  
  

From decentralized AI to DePIN to Web2, we’re seeing momentum from all sides. Builders want decentralized primitives that work like the cloud. That’s what we’re delivering.

This Month’s Drop: Akave v0.1.0

March marked a major protocol milestone: v0.1.0 is live.

This release isn’t just incremental—it’s our first stable version toward a decentralized object storage layer that actually feels like something developers would use for production workloads. Think streaming uploads, blockchain-authenticated buckets, Filecoin integration, erasure coding, and encryption. All wrapped in a clean CLI and SDK.

If you’ve ever tried to build on Web3 storage before and bounced off the UX—this version was built for you.

Why This Release Matters

Decentralized infra often means tradeoffs. Performance vs. trustlessness. Simplicity vs. modularity. Our goal with Akave is to make those tradeoffs disappear.

With v0.1.0, we’ve shipped:

🔁 Decentralized Metadata Sync

Your buckets now gossip. Literally.
We use a gossip protocol to sync bucket metadata across nodes, supporting either:

• A global namespace (shared by all)
  
  
• Or per-user segmentation (based on EVM-auth)
  
  

This gives you flexibility depending on your use case: multi-tenant, isolated, or fully public.

‍

🚀 gRPC Streaming API (with gRPC-web)

Uploading a 2GB file? Cool. We’ve got chunked streaming.
Need browser support? gRPC-web is live.
Prefer your own client-side logic? The SDK is open and composable.

You can now stream files in/out of Akave using both unary and chunked RPCs. High-performance, low-latency, no BS.

🔐 Encryption and Erasure Coding

Every chunk of data gets encrypted with HD-derived keys (yep, deterministic).
We also apply erasure coding with configurable parity blocks, so even if chunks drop, your file doesn’t.

We’re talking cloud-level durability, without giving up control.

📦 CAR File Packing + Filecoin Deals

When files hit our nodes, they get chunked, packed into CAR files, and prepped for long-term storage.
We:

• Generate Merkle proofs
• Create Filecoin deals
• Support commP-based proof generation
• PDP’s proving mechanism
  
  

All of this happens via our custom Boost client behind the scenes. You just upload—Akave handles the rest.

🛠️ CLI and SDK: Dev-First by Default

Akave is made for builders.

Our CLI supports:

• Anonymous or on-chain wallets
• Flag-based config for bucket access, encryption, chunking, and deal flow
  
  

The SDK gives even deeper control:

• Upload concurrency
• Chunk size + encryption settings
• Blockchain or streaming mode
• Retrieval fallback to Filecoin (DownloadV2 FTW)
  
  

Whether you’re automating backups, pushing AI datasets, or building DePIN systems, this stack gives you the right tools.

A Closer Look at the Architecture

‍

Akave’s v0.1.0 release spans several layers. Here’s how the stack comes together:

Node APIs

• Streaming API: unauthenticated, shared namespace
• Blockchain API: EVM-authenticated, per-user buckets
• Gossip Protocol: metadata syncing across nodes
• FlatFS: simple, fast, and proven local storage backend
  
  

Filecoin CAR Integration

• Packs chunks into CAR files
• Handles commP + proof generation
• Issues deals via Boost + tracks deal status
  
  

This is how we lock in verifiability and retrieval without making you think about sectors or sealing.

‍

SDK

• Dual-mode: streaming & blockchain
• Deterministic key derivation for metadata + chunks
• Full encryption + erasure settings
• Fallback download paths for Filecoin-native retrieval
  
  

CLI Tools

• Mirrors SDK with flag-based UX
• Works with or without wallets
• Runs in CI, local, or airgapped flows

‍

Smart Contracts

• Defines file + bucket state
• Chunks indexed on-chain
• Event logs parsed and indexed by our custom indexer
  
  

Yes, it’s fully composable. Want to integrate Akave with your own protocol or app? It’s designed to plug in, not lock you in.

What’s Next?

With v0.1.0 out the door, we’re already working on:

• Coming soon: O3: Our S3-compatible API
• Performance benchmarks for Akave Cloud
• Simplified onboarding flows for Web2 developers through our RUST SDK
  
  

We’re building for adoption—real adoption. And that means continuing to prioritize UX, developer tools, and production-grade performance.

Final Thoughts

We didn’t set out to build “decentralized storage.” We set out to build usable, verifiable, modular, and decentralized cloud infrastructure.

That’s what v0.1.0 represents: the first serious alternative to centralized object storage that works like the cloud, but gives you control like crypto promised.

And we’re just getting started.