Explore → Analyze
From screening to report.
Agents screen thousands of web3 projects and stocks, then turn any saved idea into a structured analysis report, with sources cited.
Applied AI·Built in Norway
Research agents for financial markets.
BlockMind builds research agents for stocks and digital assets: they read premium market data, apply documented analyst frameworks, and deliver verdicts with the sources and reasoning attached.
In beta · onboarding in small batches
~/blockmind/workspace
Supported by
Google Cloud
A research workspace, run by agents.
A shared digital office for the analyst and their agent: professional charts powered by TradingView, structured reports, and a research workflow that goes from screening to verdict.
Verdicts
An ensemble of analyst lenses.
Each report is scored independently against several codified analyst frameworks. Every verdict shows the reasoning and the data behind it.
Track
Continuous coverage.
Portfolio health checks, watchlists, tracking of your own sources — newsletters, YouTube, X — and a daily research brief delivered by email every morning.
Engineered end to end: data → agents → frameworks → verdict.
The product is an applied-AI system built and operated end to end by the BlockMind team — from market-data ingestion to agent orchestration to a verdict the user can audit.
Data
Premium data feeds: on-chain flows, derivatives, sentiment, news.
Agents
Autonomous, multi-step research runs powered by frontier language models.
Frameworks
Codified analyst frameworks applied as independent evaluation lenses.
Verdict
One auditable conclusion, with sources cited and the reasoning shown.
Orchestration
Durable agent workflows
Long-running, multi-step research jobs that survive restarts, retry safely, and run on a schedule — every morning, for every user.
Isolation
Hosted agent workspaces
Each user's agent runs in its own managed workspace with persistent memory — context that compounds instead of resetting per chat.
Data engineering
Market data pipelines
Data aggregated across specialist providers — Nansen, Glassnode, Coinglass, DefiLlama, LunarCrush — and normalized into agent-readable context.
Evaluation
Framework-based judging
Analyst frameworks are codified and applied as independent lenses; verdicts show how each framework weighed the same report.
Transparency
Reasoning on the record
Every output carries its sources and chain of reasoning. Users can see why — and are expected to disagree when the data warrants it.
Operations
Full-stack observability
Every agent run is traced and evaluated in production — the unglamorous work that makes AI products dependable.
Built so the human stays in charge.
BlockMind is built for markets where trust has to be earned.
Read-only
Agents can never touch funds
Wallets and exchanges connect read-only. The agent can see positions but cannot trade, withdraw, or move money — even if instructed to.
Human-in-the-loop
People decide, agents inform
Agents research, analyze, and give verdicts. They never execute. The final decision is always the user's, made on evidence they can inspect.
Transparent
No black box
Every verdict is delivered with its reasoning and sources attached. If the agent can't show its work, the verdict doesn't ship.
Analysis, not advice
Decision support, clearly framed
The platform produces research and analysis — not investment advice — and says so plainly wherever a verdict appears.