The dispatcher
A missed record can stop the work before the work even starts.
Development-stage softwareNot yet available
AI for work where the records have to hold up.
RTS.AI is building local-first document and evidence software for organizations that need clear sources, visible limits, and human approval.
Design priorities
01
Customer-controlled processing
02
Source-linked findings
03
Limits shown plainly
04
Human judgment retained
02 / Why we exist
The work is hard enough. The software should not make the evidence harder to trust.
Document-heavy work creates a quiet tax. People search for the latest record. They rebuild context that should have traveled with the file. They defend decisions with software that never showed them what was missing.
Sometimes records can move. Sometimes they cannot. Either way, the person responsible needs to know where the evidence came from, where the limit is, and who still has to decide.
The quality desk
A conclusion is only useful when the source and the gap are visible.
The records owner
The next person needs the history, not another mystery folder.
03 / Work in development
Three lanes. Three honest status labels.
These are development programs, not products available for purchase. No launch date is being promised.
Founding product · 01
RTS.AI Markets
RTS.AI Markets is the original RUN Trading Systems product. The RUN name comes from its four-run framework.
A governed market-intelligence system built to establish market conditions before examining individual tickers.
Not a trading bot, stock picker, signal service, prediction machine, recommendation engine, or investment adviser.It presents governed evidence, visible limitations, and interpretation. The user retains every decision.
- Run 1Before open · market context
Define the regime
Build a preliminary regime from the market evidence available before the open.
- Run 2After open · evidence gate
Confirm, reject, or limit
Test that context against opening behavior. Missing, stale, or conflicting evidence lowers confidence and keeps the run Data Limited.
- Run 3Sequenced review · ticker second
Review evidence objects
Examine candidates inside the market context—as evidence, never as instructions.
- Run 4Post-close · validation
Test what survived
Compare the day’s thesis with the close, preserve gaps, and carry findings into the next cycle.
Market firstEvidence before narrativeTicker secondUser decidesData Limited stays visible
Flagship development program · 02
B2B Desk
The document-heavy work deserves a better operating desk.
B2B Desk is the largest planned RTS.AI product surface—being engineered for the work behind real operations.
Engineering direction
Not another chat window. The design keeps sources, evidence, workflow, and human decisions separate—then brings them together in one governed operating desk.
Architecture / four governed separations
A polished answer should not outrun the record.
The public design direction keeps the record behind the answer visible: what supports it, what limits it, and who still has to decide.
- Approved sources
Keep identity and context attached.
The design starts from approved records and keeps their source context attached to the work.
- Evidence state
Show what supports the finding.
The intended view makes incomplete or conflicting evidence visible instead of smoothing it over.
- Governed workflow
Keep state, history, and limits in view.
Review state, preserved history, and operating limits are designed to remain visible as work moves.
- Human authority
An authorized person decides.
AI may organize, explain, and challenge. Protected actions remain behind explicit human approval.
Approved sourcesEvidence stateGoverned workflowHuman authority
Architecture direction—not released capability. B2B Desk has no live customer deployment, production connector, present-tense performance claim, security certification, or compliance result.
Supporting platform lane
Records & Filing
Designed / future work · not offered
A future records and evidence platform centered on record identity, lineage, custody state, and review history.
Live customer filing is not available. No production connector or customer deployment is being claimed.
04 / Method & boundaries
A system can carry evidence. It cannot own the decision.
The method is simple to say and hard to keep. Evidence before explanation. Explanation before judgment. When the evidence is weak, the confidence comes down with it.
- 01
Machine gathers evidence.
Sources, gaps, freshness, and conflicts belong in view.
- 02
Rules interpret governed state.
The system follows the lane’s rules instead of inventing its own.
- 03
AI translates and challenges.
It can explain, question, and propose. It cannot become final authority.
- 04
A person decides.
Protected actions remain behind an explicit human decision.
Internal build discipline
Material work is separated into builder, challenger, and owner-decision roles. Findings and dissent stay in the release record. Protected actions require explicit owner approval.
An internal method is not a certification, security guarantee, compliance claim, or proof that unreleased software performs a feature.
05 / Founder & contact
Built from the operator's side of the desk.
I have spent most of my working life inside jobs where records, handoffs, and accountability are not paperwork in the abstract. They affect whether the work moves, whether someone can defend a decision, and whether the next person has what they need.
I am a U.S. Army veteran, worked as a trucking owner-operator, and now work in flight-hardware fabrication. In each setting, I watched capable people lose time to software that was old, awkward, or built around the paperwork instead of the person doing the work.
RTS.AI grew from that frustration. I am not trying to remove judgment from those jobs. I am building tools meant to make evidence easier to trace, uncertainty harder to hide, and decisions easier for qualified people to defend.
— Adam Cornelius, founder and owner
U.S. Army veteran
Standards, handoffs, and responsibility.
Trucking owner-operator
DOT records, dispatch, and the paperwork grind.
Flight-hardware fabrication
Traceability and work that must be provably right.
RTS.AI
The work is hard. The software shouldn’t be.
Direct contact
Start with the problem, not a sales form.
Email reaches Adam directly. This page has no contact form and does not collect contact data.
rts.ai@runtradingsystems.net06 / Open source
Tools we give away.
Git Dojo
The GitHub Mastery Path — a visual, hands-on way to truly understand Git and GitHub, from first commit to expert-level workflows. Built entirely in-house at RTS.AI, best experienced on a computer, and free to use and learn from. Website ↗ · GitHub repo ↗
More to come
Future open-source software from RTS.AI will be listed here as it is released.
Support this work
Funding channels are being set up. GitHub Sponsors is pending approval, and a Stripe option is in progress. Links will appear here — and inside the apps and on GitHub — once they are live. Support is always optional; the software stays free either way.
Release notices
Privacy and accessibility.
Privacy on this page
This page does not include an analytics service, advertising tracker, contact form, account system, or RTS.AI-authored browser-storage code. The hosting and security platforms may process standard request data, IP addresses, and essential access information needed to deliver and protect the site. This notice will be updated before RTS.AI enables any new collection here.
Accessibility
RTS.AI is working to keep this page usable with keyboards, zoom, reduced motion, and assistive technology. If you encounter a barrier or need the content in another format, email rts.ai@runtradingsystems.net. Include the page and the problem you encountered so it can be reviewed.