Build Automations on the Sui stack
and its protocols. Every node remembers.
Every run generates Verifiable data.
Visual workflow builder powered by Walrus and MemWal. Drag, connect, and compose Sui protocols: DeFi, bridges, oracles, NFTs. Every execution leaves a permanent, privacy-safe dataset you can verify on-chain.
ON-CHAIN AUTOMATION, REIMAGINED
Build on Sui and its protocols.
As easily as dragging a node.
M2A is the on-chain automation builder:connect AI agents, Sui protocols, and data sources visually, without writing a single line of Move or wrangling an SDK directly. You already know how to use M2A. The difference is everything you build remembers, and everything that happens is verifiable on-chain.
Other automation tools
Stateless:every run starts from zero
Centralized:one server, one point of failure
No verification:you trust the platform's word
Off-chain only:no real connection to Web3
M2A
Persistent memory:every node remembers across runs
Built on Sui, Walrus, and MemWal:decentralized by default
Verifiable:every execution is a Sui transaction
Native protocol access:connect directly to the Sui stack
Drag
Drop a node onto the canvas:an AI agent, a Sui protocol connector, a trigger, a data source.
Connect
Draw a line between nodes. That's it:they now share memory and pass data automatically.
Run
Execute. Every action becomes a verifiable on-chain transaction. No manual wiring, no protocol-specific code.
This is what on-chain automation should have looked like from the start.
Automation platforms today run workflows.
Then the value disappears.
Without M2A
Workflow runs. Value disappears. No verifiable record. No reusable data. Every run starts from zero.
Every run.
A dataset.
Every automation run produces a verifiable, privacy-safe dataset stored on Walrus. Data compounds across runs, smarter over time.
An automation engine that gets smarter
every time you use it.
Three primitives. Workflow engine to build. Dataset generation to compound. Memory to make it smart.
Workflow
Engine
Visual canvas, 48 node types. Drag-connect-execute. On-chain by default.
Dataset
Generation
Every automation run produces a verifiable, privacy-safe dataset, automatically. This is what M2A delivers.
Persistent
Memory
The mechanism that makes automations smarter across runs and datasets more valuable.
48 node types. One canvas.
Build any automation.
DeFi
Cetus
DeepBook
Aftermath
Navi
Suilend
Haedal
Volo
Bucket
Bluefin
AlphaFi
Price Alert
Web2
Slack
Discord
Telegram
Google Sheets
Airtable
Notion
Triggers
Input
Webhook
Schedule
On-Chain Event
Form
Discord Bot
Blockchain
Sui RPC
Walrus
SuiNS
Balance Monitor
File I/O
AI Agents
M2A Agent
Spawn Agent
Code Node
Loop
Logic/Data
HTTP Request
JSON Parser
CSV Parser
RSS Reader
Conditional
Merge
Wait
Counter
Oracles
Pyth
Switchboard
Bridge
Wormhole
Sui Bridge
NFT
NFT Mint
TradePort
Floor Alert
Storage
IPFS Upload
DeFi
Cetus
DeepBook
Aftermath
Navi
Suilend
Haedal
Volo
Bucket
Bluefin
AlphaFi
Price Alert
Web2
Slack
Discord
Telegram
Google Sheets
Airtable
Notion
Triggers
Input
Webhook
Schedule
On-Chain Event
Form
Discord Bot
Blockchain
Sui RPC
Walrus
SuiNS
Balance Monitor
File I/O
AI Agents
M2A Agent
Spawn Agent
Code Node
Loop
Logic/Data
HTTP Request
JSON Parser
CSV Parser
RSS Reader
Conditional
Merge
Wait
Counter
Oracles
Pyth
Switchboard
Bridge
Wormhole
Sui Bridge
NFT
NFT Mint
TradePort
Floor Alert
Storage
IPFS Upload
DeFi
Cetus
DeepBook
Aftermath
Navi
Suilend
Haedal
Volo
Bucket
Bluefin
AlphaFi
Price Alert
Web2
Slack
Discord
Telegram
Google Sheets
Airtable
Notion
Triggers
Input
Webhook
Schedule
On-Chain Event
Form
Discord Bot
Blockchain
Sui RPC
Walrus
SuiNS
Balance Monitor
File I/O
AI Agents
M2A Agent
Spawn Agent
Code Node
Loop
Logic/Data
HTTP Request
JSON Parser
CSV Parser
RSS Reader
Conditional
Merge
Wait
Counter
Oracles
Pyth
Switchboard
Bridge
Wormhole
Sui Bridge
NFT
NFT Mint
TradePort
Floor Alert
Storage
IPFS Upload
Built on the Sui stack.
Every automation you run
produces a verifiable dataset.
Privacy-preserving pipeline: capture, redact, verify, and publish to Walrus, automatically, on every run.
Before: Raw Output
Agent responses contain PII, wallet addresses, personal data:all exposed, all unverifiable.
After: Clean Dataset
Every run produces a verifiable, privacy-safe dataset stored on Walrus. Data compounds across runs, smarter over time.