Roadmap
What comes next.
What we are building, what comes next and what we are still studying. There are no dates on this page. The order reflects priority, and an item moves to the changelog the day it ships. Everything listed here is in development and described that way.
- Dates · None. Priority only.
- When it ships · It moves to the changelog.
- Order · Moves with what services ask for.
Next 9
Recently shipped
- Feedback straight to VERA Forensics
- 1.16.37 · Seal processing Indexing is much faster on large seals: mappings are processed in batches, and traces are written several times faster into a case that already holds millions of them.
- 1.16.36 · Seal processing An archive in which one file cannot be written whole, or whose checksum differs from the one the archive records, no longer fails the whole seal. The file is kept as the archive holds it, a warning names it in the logs, and processing carries on with the rest without stopping. Only a full disk stops the extraction.
- 1.16.36 · Seal processing A seal that produces tens of thousands of mappings, such as a complete phone, no longer freezes the window when the case opens or during processing. A case that froze each time it was reopened opens again.
Analysis
Cross-seal analysis Next See what links the seals of one case without reading each one first.
What it changes for the examiner
- Automatic overlaps between seals: the same number, account or picture on two devices.
- Traces present on one seal and deleted on another, flagged for review.
- Duplicates across seals merged, with each original kept as the source.
Already today. The assistant already answers questions across the seals of one case, one sub-agent per seal.
People and entities Next 1 update shipped One person, whatever the application they appear in.
What it changes for the examiner
- Contacts from every application merged into people and organisations, each merge confirmed by the examiner.
- A summary sheet per person: contacts, places, activity over time, every line cited.
- Behavioural profiles are being studied, as proposals the examiner validates, never as conclusions.
Framework. Under legal authority, with every statement traced back to its source. Responsible use
Seal dashboards Next An overview of each kind of activity, built by the assistant and kept up to date.
What it changes for the examiner
- Web history: categories of sites visited, the share of each browser, how habits change over time.
- Messaging: main contacts, recurring topics, who supplies and who buys.
- Accounting files: companies that only appear in invoices, amounts out of pattern.
- Clickable charts, trees and graphs, filtered by period, each one opening the traces behind it.
Media classification and search Next Sort and search thousands of pictures and videos by what they show.
What it changes for the examiner
- Automatic categories for pictures and videos, which the examiner can correct.
- Text read from pictures and scanned documents, searchable like any message.
- Near-identical pictures grouped, so a burst of photos reads as one event.
- Face matching is being studied, only where the law allows it.
Framework. Face matching would only be available where the applicable law permits it, under legal authority. Responsible use
Relationship graphs Exploring Who talks to whom, how often and about what, on one screen.
What it changes for the examiner
- Graphs that start from what matters most and hide the rest until asked.
- Links weighted by volume, period and topic, each one opening the messages behind it.
- Large graphs that can be explored in depth, including in three dimensions.
Deeper views Exploring 1 update shipped Group and link records inside every view.
What it changes for the examiner
- Group, pin and link records without leaving the view.
- Travel tickets and bookings read as journeys.
- The content of a visited web page shown next to the history entry, when the service allows a connection.
OSINT and ADINT Exploring Cross-check identifiers found on the device against outside sources.
What it changes for the examiner
- Usernames, numbers and accounts found on the device checked against open sources (OSINT).
- Advertising identifiers present on the device used to link accounts and devices (ADINT).
- Only through a connection the service turns on, never by default.
Framework. Under legal authority, within the framework that applies to the investigation. Responsible use
Assistant and AI
A more autonomous assistant In progress 1 update shipped Long tasks that carry on, and a clear view of what is covered.
What it changes for the examiner
- Background tasks resume after VERA is closed and reopened.
- Sub-agents that start a mapping or a summary of a whole seal themselves, after the examiner's yes.
- A coverage table per seal: installed, holds data, mapped, indexed, in the library.
Exact counts and cohorts In progress The assistant counts over the whole case, not on a sample, and keeps its method.
What it changes for the examiner
- Questions such as “how many”, “which ones”, “after each call” or “never”, answered over every app and seal of a case.
- Each calculation keeps its exact method, can be run again and stays listed with the case.
- The traces a calculation finds become a cohort: open it in the search, filter any view, label it in one step, export it with its method.
- Long calculations run in the background, with progress and a stop button.
An AI pass over every message Next 2 updates shipped Read every message once, and bring out what departs from the ordinary.
What it changes for the examiner
- Every message read during seal processing, with anything out of the ordinary recorded in a structured list.
- A second pass over what the first one raised.
- Search inside what the pass produced, next to the original text.
- Filters such as spam, newsletter or suspect, and hierarchical labels.
- Every finding links to the message it came from.
Already today. Signals already flag traces that look like suspicious behavior, and an optional AI review says whether the passage really looks suspicious, whether the words are ordinary, or whether it cannot tell.
Already shipped
- 1.16.23 In Messages, the Suspicion button sorts conversations from most to least suspect by their classification, or keeps only the suspect ones, from 50, 70 or 90%. Each choice shows how many threads it keeps, and a suspect tab shows its percentage.
- 1.16.16 AI review: an optional step reads the flagged traces with the messages around them and says whether the passage really looks suspicious, whether the words are ordinary, or whether it cannot tell. Weak signals are read too. The estimate (time on the workstation, cost in euros at VERA Forensics) appears before it starts.
Reports and dynamic exports Next 4 updates shipped A report a magistrate can read, with sources one click away.
What it changes for the examiner
- Configurable reports in PDF, Word, OpenDocument and HTML, and tables in CSV or Excel.
- Interactive HTML reports: filter by period, every source clickable.
- A short, clear synthesis up front, the detail in annexes.
- Drafting with the assistant, revised over several passes, including from an existing report.
Already today. Labelled traces already export to PDF, Word and HTML.
Already shipped
- 1.16.31 Excel joins PDF, Word and HTML wherever they were offered, and every view in Table mode, Generic data included, exports to Excel. The workbook is the table on screen with all matching rows: each header filters and sorts, dates are real dates in the time zone on screen, phone numbers stay text, and each label has its own column.
- 1.16.31 From a Messages thread, the Excel export takes this thread, all conversations or the search results. Each message carries its conversation, so filtering one conversation and then one label gives its key messages.
- 1.16.31 The Map and the Timeline export what is framed on screen, with its zoom and filters, to PDF, Word, HTML and Excel. The HTML page is a single file that opens without a network, with zoom, hover, details and sources; on the Map, a click on a cluster lists its traces.
- 1.15.183 Exports read better: in PDF the time fits on one line, in Word consecutive messages are separated, in HTML metadata reads field by field, and group members are listed once at the top of a thread.
Coverage and mapping
More reliable automatic mapping In progress 2 updates shipped Better first attempts on applications nobody has mapped yet.
What it changes for the examiner
- Examples drawn from the library according to the kind of application being mapped.
- Work split across sub-agents for large applications.
- The difference between two attempts shown on the mapping graph.
- Dedicated test sets, measured on every change.
Already shipped
- 1.15.193 Running automatic mapping again on a seal that was already processed is faster: a database that did not change is neither read nor sent to the AI a second time. The first run is still the longest.
- 1.15.190 Automatic mapping now learns from the public library: for each view, it works from complete, validated examples taken from it.
Processing under time pressure Next 2 updates shipped Start working on a seal before everything is finished.
What it changes for the examiner
- A library-only pass that makes known applications readable first.
- The examiner works on the seal while the AI keeps adding mappings in the background.
- Meaning-based search and conversation summaries deferred or run in parallel.
Already shipped
- 1.15.190 Seal test mode samples 100 traces per mapping. A mapping already indexed is not read again: its index is kept, and marked out of date if it no longer matches the settings.
- 1.15.183 Seal import, test mode: an advanced switch, off by default, indexes only a random sample of each mapping to try a large seal without waiting.
Map everything, beyond the phone Next 1 update shipped Computers, logs and any readable file, through the same mappings.
What it changes for the examiner
- Computer sources on Windows, macOS and Linux: system traces, browser profiles, desktop messaging, documents and mail exports.
- Plain-text log files and application journals.
- Exports already produced by other tools, taken as input.
- Any readable file becomes a source that a mapping turns into a view.
Already today. Phone extractions are already covered, including files that are not databases.
Library and traceability
Traceability for the service Next Read back how each conclusion was reached.
What it changes for the examiner
- Full conversations between an examiner and the assistant, easy for the service to read back.
- A log of what entered and left the service library, and who approved it.
Teams
Working on a case together Exploring 1 update shipped Several examiners of one service on the same case.
What it changes for the examiner
- The service server holds the shared state of a case.
- Mappings, labels, assistant conversations and their summaries shared between colleagues.
- Evidence never leaves the service.
Feedback straight to VERA Forensics Shipped in 1.15.183 Send a remark from inside VERA, with the logs if you choose.
What it changes for the examiner
- The feedback button sends the message to VERA Forensics by email when the VERA Internet connection is on.
- A reply address and the recent logs can be attached.
Missing a format, a view, a question?
The order on this page moves with what services tell us. If something your laboratory needs is not here, say so.
Write to VERA ForensicsSee it running on a real extraction
A service asks for a demonstration, usually thirty minutes on fictitious data. An independent expert asks for access: a person reviews the request, then the yearly licence is ordered from the account.
Every request is reviewed before access is granted.