Bloom Terminal

Onchain agents that can trade, loop, swap, harvest, rebalance, and exit across live markets.

[Bloom Terminal @usebloom_base]

Bloom dApp preview.

An on-chain agent wallet built to run DeFi strategies automatically.

Loop, swap, leverage, harvest, rebalance, and exit through one runtime.

The user sets the strategy.
The agent handles execution.

DeFi is becoming agent-run.

Bloom is building the layer for it.

[Bloom Terminal @usebloom_base]

https://t.co/CZmtNQzG5L

[Bloom Terminal @usebloom_base]

Bloom AI Council is now live.

Most AI-powered DeFi apps rely on a single model to make recommendations.

Bloom doesn't.

Every recommendation is reviewed by a council of specialized agents:

→ Yield Analyst
→ Risk Analyst
→ Smart Money Analyst
→ Market Analyst
When a user asks:
"Where should I deploy 500 USDC?"
the process looks like this:

User asks Bloom
→ Bloom scans available opportunities
→ Yield Analyst evaluates returns
→ Risk Analyst evaluates downside
→ Smart Money Analyst analyzes capital flows
→ Market Analyst reviews broader conditions
→ Agents debate independently
→ Consensus score is calculated
→ Confidence score is generated
→ Final recommendation is returned

Users receive:

• Recommended vault
• APY
• Risk assessment
• Council consensus
• Confidence score
• Simulation access
• One-click deployment through Agent Wallets

This is a major step toward agent-driven capital allocation.
Not one AI opinion.
Multiple agents.
Independent analysis.
Consensus-based recommendations.

Built on @base .

[Hatto Hanzō @HanzoYasunaga]

$bloom and @IBM

👀

[Bloom Terminal @usebloom_base]

Another step forward.

Bloom is joining the IBM Partner Plus ecosystem.

As we build the capital intelligence layer for DeFi, we're excited to learn, grow, and accelerate alongside some of the world's leading technology organizations.

Thank you @IBMPartners @IBM.

Back to shipping.

[Bloom Terminal @usebloom_base]

nice article

[metape @metapee]

https://t.co/VxpVB5f71W

[Bloom Terminal @usebloom_base]

We've been shipping non-stop.

Over the last few weeks Bloom has evolved from a simple DeFi interface into an AI-powered capital intelligence platform.

Recently shipped:

• Smart Money Analysis
• Bloom Simulations
• Digital Twin Framework
• Privacy Layer
• Copilot Actions
• Agent Council
• Agent Reputation
• Outcome Tracking

Now Bloom can:

→ Analyze opportunities
→ Simulate thousands of scenarios
→ Debate decisions across multiple AI agents
→ Track agent performance over time
→ Measure outcomes
→ Protect strategy intelligence

The vision is simple:

Observe.
Simulate.
Protect.
Act.

We're just getting started.

[Bloom Terminal @usebloom_base]

New in Bloom

Smart Money Analysis.

Bloom continuously analyzes Base market activity and ranks opportunities using:

• Liquidity
• Volume
• Turnover
• Momentum

See something interesting?

Click "Explain".

Copilot researches the opportunity, analyzes the signal, and explains why it was ranked.

No endless dashboards.

No hunting through data.

Just ask Bloom.

Observe.
Simulate.
Protect.
Act.

[Bloom Terminal @usebloom_base]

you can test it now : https://t.co/DjzX7k5ZbF

[Bloom Terminal @usebloom_base]

Introducing Bloom Privacy Layer

Your strategy is your edge.

Privacy Layer v0.1 helps protect the intelligence behind your decisions:

✓ Private Mode
✓ Hidden amounts
✓ Hidden wallet addresses
✓ Hidden strategy labels
✓ Encrypted Copilot memory
✓ Encrypted simulation history
Bloom now protects strategy metadata while keeping users fully in control of their capital.

This is the first step toward a much bigger vision:
Private intelligence for DeFi.

🔒 Your strategy.
🔒 Your data.

built on @base

[Bloom Terminal @usebloom_base]

Dev Update

Today we shipped Bloom Simulations.

Before capital moves, Bloom can now model up to 10,000 possible futures and evaluate how a position performs across different market conditions.

Examples:

• BTC +20%
• BTC -30%
• Stablecoin depeg
• Liquidity crunch
• Base ecosystem expansion
• APY compression

Behind the scenes, Bloom combines:

• Macro Agent
• Yield Agent
• Risk Agent
• Market Agent

to stress test positions and identify risk before deployment.

Instead of only showing:
APY = 12%

Bloom can now estimate:

✓ Expected APY
✓ Survival Probability
✓ Max Drawdown
✓ Liquidation Risk
✓ Risk-Adjusted Outcome
The goal isn't to find the highest yield.

The goal is to find the best outcome after simulating thousands of possible futures.
This is the first step toward our long-term vision:

A Financial Digital Twin of the on-chain economy.

Before capital moves.

The future is simulated.

TRY : https://t.co/iYjRJUaZWn

[Hatto Hanzō @HanzoYasunaga]

$bloom this is very good tech

[Bloom Terminal @usebloom_base]

Meet Bloom Copilot.

Your portfolio.
Your memory.
Your market terminal.
Ask:
→ What should I do today?
→ Show ETH chart
→ Analyze VVV
→ What are smart wallets buying?
→ Deposit into the best vault
→ Withdraw my position

Powered by MiMo reasoning.

Grounded in live market data, portfolio data, and agent memory.
Every action stays execution-safe and user-approved.

Talk to your capital.

Copilot : https://t.co/iYjRJUaZWn

[Bloom Terminal @usebloom_base]

Today, Copilot took another step forward.

Instead of navigating through dashboards and forms, users can interact with Bloom using natural language.

Ask.
Confirm.
Execute.

The future of DeFi isn't more buttons.

It's conversation.

Try copilot : https://t.co/bHZslVN8Fi

[Bloom Terminal @usebloom_base]

Introducing Bloom Copilot.

An AI-powered DeFi assistant built directly into Bloom.

Powered by MiMo reasoning and Bloom's execution-safe infrastructure, Copilot helps users understand, monitor, and interact with their capital through natural language.

You can ask:

→ "How is my position performing?"
→ "What are the risks?"
→ "Are there better opportunities?"
→ "What changed since my last deposit?"
→ "Deposit 10 USDC into the best vault"
→ "Withdraw my position"

Bloom analyzes your portfolio, reviews opportunities, compares vaults, and prepares actions directly from the chat interface.

Every action is validated by Bloom.
Every transaction remains user-approved and wallet-signed.

No endless dashboards.
No hunting through menus.
Just talk to your capital.

Bloom Copilot is now live.

Check here : https://t.co/iYjRJUbxLV

[Bloom Terminal @usebloom_base]

Agent Memory v0.1 is now live in production and represents an important milestone in Bloom's evolution from a yield deployment interface into an intelligent agent platform.

What was released:

• Agent Memory is now available within Position Details for every Bloom agent.
• Agents now maintain a local financial memory layer based on their historical activity.
• Deposits and withdrawals performed through the Agent Wallet flow are automatically recorded after confirmed on-chain execution.
• Memory persists across sessions and tracks historical observations over time.

Current memory categories include:

1. Identity

* Agent address
* Creation timestamp
* Last active timestamp

2. Capital

* Total deposited
* Current position value
* Highest position value observed
* Estimated yield

3. Yield

* Current APY
* Highest APY observed
* Lowest APY observed
* APY snapshots

4. Vault History

* Current vault
* Previous vault
* Deployment count
* Withdrawal count

5. Activity

* Last action
* Last transaction
* Activity timestamps

Why this matters:

Most DeFi applications only show the current state of a position.

Bloom is beginning to give agents historical context.

An agent that remembers is fundamentally more useful than an agent that only executes.

Agent Memory creates the foundation for future intelligence layers including:

• Yield Recommendations
• Opportunity Detection
• Historical Performance Analysis
• Agent Decision Systems
• Autonomous Execution Frameworks

Without memory, an agent cannot understand what happened previously.

With memory, an agent can begin to reason about performance, identify changes, and generate informed recommendations.

Strategic Impact:

This release moves Bloom one step closer to becoming an intelligence layer for on-chain capital.

The progression is now:

Agent Wallet
→ Agent Memory
→ Yield Intelligence
→ Recommendations
→ Autonomy

This is an important distinction for our positioning.

We are not simply building another yield dashboard.

We are building financial agents that accumulate context over time and become increasingly useful as they interact with capital.

Current Status:

✓ Live in Production
✓ Agent Wallet Integrated
✓ Shielded Yield Beta Live
✓ Deposit Recording Active
✓ Withdrawal Recording Active
✓ Historical Memory Persistence Active

Agent Memory v0.1 is the first step toward creating agents that do more than execute transactions.

It is the beginning of creating agents that learn from their own financial history.

[Bloom Terminal @usebloom_base]

Bloom Update

Agent Memory v0.1 is now live.

Most agents today can execute.

Very few can remember.

Every Bloom agent now builds a financial memory layer from its own activity:

• Capital deployed
• Position value
• Yield history
• Vault history
• Activity history
• APY observations
This creates context for future intelligence.
An agent shouldn't just know where your capital is.
It should understand where it's been.
Alongside Agent Memory, we've also shipped:
✓ Agent Wallet Execution (Beta)
✓ Shielded Yield (Beta)
✓ Morpho Integration
✓ Deposit via Agent Wallet
✓ Withdraw via Agent Wallet

Current flow:

Main Wallet
→ Agent Wallet
→ Yield Strategy
→ Withdraw

Today, Agent Memory is read-only and stored locally.

Tomorrow, it becomes the foundation for recommendations, yield intelligence, and smarter agents.

Agents need memory.
Financial agents need financial memory.
Bloom.

[Bloom Terminal @usebloom_base]

Bloom Intelligence CLI Engine is now live

Install:

npm install -g @bloomterminal/cli
Analyze Base protocols:
bloom intel

Analyze Morpho opportunities:

bloom morpho
Analyze a wallet:
bloom wallet 0x...

Generate a report:

bloom report 0x...
Export research:
bloom export-report 0x... --format markdown

The flow is simple:

wallet
→ intelligence
→ risk analysis
→ opportunity discovery
→ report generation
No signing.
No custody.

No execution.

Capital needs more than dashboards.
It needs intelligence.

[Bloom Terminal @usebloom_base]

Bloom Intelligence Engine is becoming the research layer for on-chain capital.

Most platforms show data.

Bloom is focused on helping users understand what that data means.

Today users can:

monitor protocol growth across Base

track Morpho vault flows, utilization, and APY trends

identify portfolio concentration and idle capital

discover yield opportunities

generate weekly intelligence reports

build private research through the Shielded Notebook powered by @Veilnet_

The goal is simple:

understand where capital is moving

identify opportunities

manage risk

build conviction

without jumping between multiple dashboards and tools.

Data is the starting point.

Intelligence is the destination.

This is one step toward a future where intelligence, private research, portfolio context, and AI agents work together inside a single platform.

Built on @base .

[Bloom Terminal @usebloom_base]

Development Update:

Bloom Intelligence Engine

The goal is simple:

help users understand where capital is moving before they decide where to move theirs.

The first version includes:

Protocol Radar

Track protocol growth across Base, including TVL trends and ecosystem activity.

Morpho Intelligence

Monitor vault performance, APY trends, utilization, capital flows, and yield opportunities.

Base Opportunity Feed

Surface unusual activity, liquidity changes, volume expansion, and emerging opportunities across the Base ecosystem.

Wallet Copilot

Analyze portfolio composition, concentration risk, idle capital, and potential yield opportunities.

Most dashboards stop at information.

Bloom is moving toward intelligence.

Not just:

"what happened?"

But:

"what does it mean?"

and

"what should I pay attention to next?"

No execution.
No custody.
No transactions.
Just intelligence.

This is another step toward Bloom's long-term vision:

Intelligence
→ Agent Reasoning
→ Private Knowledge
→ Portfolio Context
→ Autonomous Execution

One step at a time.

[Bloom Terminal @usebloom_base]

@bdiasti_ai Appreciate it 🤝

Yield was the starting point.

The bigger vision is an ecosystem where intelligence, agents, portfolio context, and private knowledge work together instead of existing across dozens of disconnected tools.

Still building.

[Bloom Terminal @usebloom_base]

Bloom Ecosystem so far.

What started as a yield terminal is expanding into a broader agent and intelligence stack across DeFi, privacy, automation, and portfolio management.

Here is what users can use today:

Agent Wallet

An isolated user-controlled wallet that separates execution from the main wallet, allowing users to manage capital through a dedicated agent environment.

Stable Yield Agent

Deploy capital into curated Morpho vaults through Bloom while maintaining a non-custodial architecture and direct user control over funds.

Shielded Notebook powered by @Veilnet_

End-to-end encrypted notes, research, agent journals, and strategy documents.

Encryption happens client-side through VeilNet's encrypted data infrastructure.

Bloom stores ciphertext only.

Bloom cannot read notebook contents.

Runtime & Skills

Agent infrastructure designed to support permissions, scheduling, recommendations, memory, and future autonomous workflows.

BLOOM Access Layer

Token-gated access that aligns platform usage with ecosystem participation while unlocking additional agent capacity.

Portfolio Context

Unified visibility across wallets, deployed capital, agent activity, and positions, giving users a single operational view of their capital.

The benefit is simple:

Users can manage capital, organize research, deploy yield strategies, and prepare agent workflows from one place.

Research becomes structured.

Knowledge becomes portable.

Capital becomes easier to manage.

Agent context becomes more useful.

Bloom is no longer just a yield interface.

It is becoming the foundation for intelligent capital infrastructure.

The long-term vision:

Intelligence
→ Agent reasoning
→ Private knowledge
→ Portfolio context
→ Autonomous execution

One step at a time.