Capture
Save the source and the reason it matters.
BrainBug is built to preserve the link, the operator context, the question that caused the capture, and the follow-up thread it belongs to.
Private Research Brain
BrainBug keeps claims from getting slippery.
A private research brain for source context, uncertainty, contradictions, freshness, and the follow-up questions future agents need.
Research usually fails after the first summary. The link gets buried. The caveat gets shaved off. The next decision inherits confidence without the receipt.
BrainBug exists to stop that. It keeps the source context close to the claim, the claim close to its label, and the label close to the decision it might influence.
How It Works
A research system with a spine.
BrainBug is built around a simple operating bet: useful agents need memory, but memory is dangerous unless it keeps provenance, uncertainty, and contradiction attached.
Distill
Break research into claims, not vibes.
A useful research brain does not store one giant summary and call it knowledge. It separates facts, opinions, theories, predictions, caveats, and open questions.
Challenge
Keep contradiction visible.
If a claim conflicts with another source, goes stale, or only has weak support, the system should say so before an agent turns it into strategy or copy.
Return
Give evidence back to the agent.
BrainBug does not publish the page. It returns source context, uncertainty, and next-step cues so the acting agent can make better decisions.
Claim Labels
The label is part of the meaning.
A fact, an expert opinion, a useful theory, and a stale claim can all be valuable. They are not the same thing. BrainBug's job is to keep that distinction visible when the work gets fast.
That matters for SEO, page strategy, reports, App Writer contracts, and any research trail where a confident sentence can become an expensive decision.
Fact
A source-backed statement that can be checked against a citation, date, quote, or observed artifact.
Opinion
A judgment or interpretation. Useful, but not allowed to pretend it is a measured fact.
Theory
A working model worth exploring. It can guide tests, but it still needs stronger evidence.
Stale
A claim that may have been true when captured but should be rechecked before it carries weight.
Contradicted
A claim with meaningful disagreement in the source trail. It gets surfaced, not hidden.
Product Personality
Personality shows up as behavior.
For BrainBug, personality is not a cute voice layered over the app. It is the operating posture: skeptical, source-aware, and loyal to the operator's actual question.
Skeptical by default
The tone is not suspicious for sport. It asks whether the claim has enough support for the job it is being asked to do.
Operator-loyal
It keeps the original question, business pressure, and decision context attached so later work does not drift from the reason the source was captured.
Receipt-first
A claim without provenance is treated as a lead, not an asset. The system wants the quote, URL, date, and source role close to the claim.
Where It Fits
The hidden layer behind stronger public work.
The public output should be normal, crawlable HTML. The private system behind it should keep source context, review status, and caveats attached so public pages do not turn guesses into confident copy.
Useful For
- SEO and AI-search research trails
- Article and page strategy
- App Writer handoffs
- Source-backed reports
- Product doctrine decisions
- Competitor and SERP investigations
- Follow-up questions for future agents
- Claims that need receipts before money moves
Refuses To Be
- Not a loose notes folder.
- Not a chatbot memory dump.
- Not an automatic publisher.
- Not a place to launder guesses into certainty.
Evidence Before Output
The point is not to remember more. The point is to trust less blindly.
BrainBug is for operators who want agents to verify claims, track sources, preserve uncertainty, and come back later without losing the thread.