The application is not publicly available and has no announced launch date. All content can be read without an account or registration details.
Illustration of a development direction, not an operational application interface.
Multiple accounts require consistent financial records
Records spread across accounts, e-wallets and cards add data entry and make errors harder to see.
Why is the ledger higher than the statement?
Missing: supermarket+350,000
Duplicate: coffee−200,000
Mismatched: electricity+72,000
Unknown: “Digital service”+99,000
Ledger minus statementVND +321,000
Illustration with educational figures from the guide Statement reconciliation. Not user data or application output.
Manual data entry is time-consuming
Gathered and re-entered from several sources each period.
Missing, duplicate or mismatched records
They distort balances and reports.
Internal transfers misclassified as expenses
Moving money between your own accounts is not spending.
Fees and recurring payments are difficult to track
Easy to miss without a periodic review.
AI grounded in your finances, not making decisions for you.
Existing foundation — not publicly available
Source code exists; complete user workflows and end-to-end verification are not established.
Existing foundation — not publicly available
Source code exists; not verified end to end.
Partial foundation — not publicly available
Some processing exists; no complete workflow.
Development direction
Not implemented or verified.
Concept under consideration
Not selected; not a commitment.
Import
Existing foundation — not publicly available
Data import from Excel files
Import transactions with a configured format. No PDF, every statement format or complete n8n connection.
Record
Existing foundation — not publicly available
Income, expenses and multiple accounts
Record and retrieve transactions across accounts. No automatic bank connection or guaranteed real-time balance.
Existing foundation — not publicly available
Recurring transactions and planned entries
Track upcoming obligations. Projections come from entered data, not AI predictions.
Plan
Partial foundation — not publicly available
Budgets and overspending alerts
A server-side foundation exists without complete user workflows. Email, push and automated alerts have not been verified as operational.
Partial foundation — not publicly available
Spending plans
Builds on recurring and planned entries. A comprehensive spending-planning workflow has not been implemented.
Illustration of data relationships between areas; dashed outline: partial.
Development direction
AI-native: financial context, bounded assistance
Make AI agents part of financial workflows, not just add a chat box — proposed design principles, not operational capabilities.
Illustration of a proposed principle: only you approve ledger changes.
Context with sources you can check
Uses only permitted data; each finding names its sources and missing data rather than inferring a confirmed balance.
Agents with a defined scope
Finds missing or duplicate entries and suggests categories with permitted tools; asks for review instead of expanding its own permissions.
Explain proposals, approve changes
Evidence, reasoning and uncertainty so you can approve, amend or reject. Read access is not permission to write.
Ledger
Written only after your approval.
⊘ Outside agent scope
Payments
Transfers
Investment trades
Asset custody
Automatic payments, transfers, investment trades and asset custody are not part of this direction. AI has not been implemented or assessed for accuracy; this is not a safety or investment-return guarantee.
Development direction
Development direction: reconciliation and automation oversight
The first priority is data quality: reconcile the ledger with statements, with evidence for your confirmation.
Development illustration — not an operational feature.
Period statement
The source you supply.
Compare with the ledger
Same account, same period.
Review list
Each entry with its reason.
Missing
On the statement, not in the ledger.
Duplicate
Recorded twice.
Mismatched
Different amount or date.
You decide
Retain, adjust or disregard.
Development direction
Statement reconciliation
A review list with evidence. Missing data does not prove a transaction does not exist.
Development direction
Discrepancy checks for automated workflows
Alerts only with a reconciliation source and defined period; incomplete data cannot establish accuracy, and issues may be missed.
Development direction
Transactions requiring verification
Fee increases, repeated or unusual spending — prompts to verify, not findings of fraud.
Additional development directions
TiHoMo ledger
Development direction
MCP for AI agents
Query data within granted permissions. No MCP server yet.
Development direction
Events for n8n and automated workflows
Events on transactions, discrepancies or overruns. No public integration.
Development direction
Third-party financial record integrations
Connect tools such as MISA to reduce re-entry. No integration exists; no partnership is confirmed.
Development direction
Gold, cryptoasset and securities tracking
Cost basis, portfolio and profit or loss. No order placement or asset custody.
Development direction
Custom reports
By period, account or category. No report builder yet.
Concept under consideration
Concepts under consideration
Research priority: a data-review list, then category suggestions and recurring-item reviews, including fee increases. Budget scenarios only after reliable data; savings goals remain an idea.
Illustration: dashed links are proposed connections with no release schedule.
Development direction
Product vision and the evidence still to build
How each hypothesis could be evaluated before further building, not results or commitments.
Proposed distinction — a value hypothesis, not a proven competitive advantage.
Chat only
Supervised agent (chosen)
Autonomous execution (excluded)
How it works
General answers, separate from records
Sourced context → proposal → your approval
AI writes records and pays on its own
Trade-off
Easy to use; the review work remains
Needs permission design and review effort
Hard-to-reverse errors; beyond user control
Illustration of evidence gates in order: a gate opens only after the previous one passes. No dates or commitments.
Validate data quality
Reconcile authorised, labelled data; report scope and limits before any accuracy claim.
Test category suggestions and recurring reviews against manual work; ledger changes only within approved scope.
Measure:AcceptedCorrectedRejectedReview time
Test source-grounded budget scenarios
Only after reliable data: scenarios from recurring obligations and explicit assumptions, not predictions or investment advice.
Measure:Understands assumptionsCompletes a plan
This site reports no measured results for these AI hypotheses, market-demand evidence or growth figures. The milestones have no committed dates; an approved data policy is needed before trials with users’ financial data.
Frequently asked questions and transparency
Has TiHoMo been released?
No. The application is in development, is not publicly available and has no announced launch date. This site describes the context, existing source-code foundations and development directions under consideration.
Does “Existing foundation” mean the application is ready to use?
No. This label confirms that source code exists for an area, not end-to-end verification or availability to external users. It is therefore always accompanied by “not publicly available”.
Does “Development direction” represent a release commitment?
No. These are proposed capabilities without committed implementation dates or scope. Development directions may change or may not be implemented.
Has AI-assisted statement reconciliation been implemented or validated?
No. AI reconciliation and the agent capabilities described here have not been implemented or validated, so no accuracy assessment is available. TiHoMo does not claim perfect accuracy. The proposed approach presents entries requiring review with supporting reasons for user confirmation, rather than modifying records on the user’s behalf.
Does TiHoMo connect to banks, MISA or n8n?
Not currently. Source code exists for data import from Excel files. Automatic connections to banks, third-party financial record tools or automation tools remain development directions.
Does this site collect personal information?
No. The site has no forms, does not require an account or email address, and does not use analytics, advertising or tracking cookies.
Can I register for launch notifications?
No registration or contact channel has been opened for launch notifications. Details will be provided here when an official channel is available.
How will the application process financial data?
No data policy has been approved for publication. This site therefore makes no commitments about how the application will process data. Do not provide statements or bank login details based on this introductory material.
Do the guides explain how to use TiHoMo?
No. The articles explain manual methods that can be applied with a ledger, spreadsheet or any application. Each article includes its own references and update date.
What does AI-native mean for TiHoMo?
It is a design direction that makes AI part of financial workflows: use source-grounded context, perform tasks through bounded tools and explain proposals for user approval. It means more than a chat box or an additional AI integration, and does not mean TiHoMo already has operational AI.
Can AI agents change records, transfer money or trade investments on their own?
Not under this proposed direction. Permission to read or analyse would not permit independent ledger changes: users would review intended changes and evidence, then approve, amend or reject them before they are recorded. Payments, transfers, investment orders and asset custody are outside the proposed agent scope. This approval mechanism has not been implemented or validated either.
Explore AI assistance under your control
Explore TiHoMo’s AI-native direction or start with practical methods for managing finances across accounts. All content can be read without registration.