Agentic workflows
AI agents that plan, use tools and verify their own work, with a person approving what matters.
Exploring
Independent AI Lab
Thinking,
Thinking,
amplified.
smind.ai explores how people and intelligent systems reason, build and decide together — through real projects, transparent costs and open notes.
Vision
Intelligence is becoming abundant. Judgment is still rare. We build tools and workflows where AI does the heavy lifting and people keep the final say.
Focus areas
Where we are exploring.
Three directions we study through hands-on work. Each one starts as an experiment and ships only once it proves useful.
Knowledge & memory
Systems that remember context, learn from corrections and turn scattered notes into usable knowledge.
ExploringLean AI infrastructure
Serverless, cost-aware building blocks so small projects can run on AI and still pay for themselves.
ExploringApproach
Plan. Act. Verify.
Keep what works.
Every idea goes through the same loop before it becomes anything public.
- 01Start from a real problemReal constraints and real budgets.
- 02Let AI do the heavy liftingAgents draft, search, build and test.
- 03Verify against realityA result counts only after it is checked, not because a command exited cleanly.
- 04Share the notesWhat worked, what failed and what it cost.
run · illustrativelive
planbreak the goal into checkable steps0.4s
searchgather current, cited sources1.2s
actcall tools, write and run code3.8s
verifytest the result against the goal0.9s
reviewhuman approves what ships—
steps 5/5status: verified
Principles
Small, honest, durable.
The rules we hold ourselves to.
01
Human in the loop
People make the consequential calls.
02
Evidence over hype
Claims are backed by tests, numbers and sources.
03
Cost-transparent
Every build states what it costs to run.
04
Privacy by default
No tracking cookies. No data resale.
Contact
Something thoughtful
is on the way.
For collaboration, partnerships or the domain itself, send a note.