Choose one or several standards, models or methodologies. VirteVal turns them into assessment rules, reads your documents and repositories, and reports for every rule what is met, what is missing and how to fix it, citing the exact passage behind each verdict.
Teams interpret long standards by hand, collect evidence in folders and redo the work when a second framework appears. Conclusions are hard to reproduce, and gaps surface late, usually just before an audit.
Today
Someone reads a 100-page standard and builds a checklist in a spreadsheet, with their own interpretation.
With VirteVal
Rules are generated from the source text, each one quoting the statement it comes from, and approved by people.
Today
"We think we comply", with nothing that points to where the proof is.
With VirteVal
Every verdict cites the document passage or repository file it is based on, or says the evidence is insufficient.
Today
Adopting ISO 29110, SAMM and Scrum means three separate assessments of the same project.
With VirteVal
One assessment covers several frameworks, and crosswalks let one piece of evidence count wherever it applies.
Every screen below is the real platform, assessing a fictional project, Atlas ERP, built from six documents and a small code repository.
Answer whether the software is critical, what you want to achieve and how many teams build it. VirteVal recommends frameworks and proven combinations, and adopts them in a project with one click.
Upload a standard, methodology or internal policy and VirteVal extracts its normative statements into rules. Each rule keeps a literal quote of its source, its subject, action and object, applicability conditions and lifecycle stage.
Your own framework becomes assessable without building checklists by hand.
Documents in their native formats are read and split into fragments: PDF, Word, PowerPoint, Excel, Markdown, images (with OCR and vision models) and diagrams. Each rule is judged element by element, and every verdict links to the fragment it quotes.
Findings become recommendations, action items and a gap report per framework, with the criterion, the document concerned and its status. After each new evaluation, corrective actions are closed or reopened according to the evidence.
Connect a Git repository or upload a ZIP. Deterministic checks, and optionally an AI code review, evaluate the code against your frameworks. The architecture is extracted from Compose, Dockerfiles, Kubernetes, Terraform, OpenAPI, ORM models and imports.
VirteVal derives the artifacts each framework requires and builds templates from the rules, merging equivalent artifacts across frameworks and flagging conflicts. Download them as Word, or let the platform draft them from your project and repository facts.
The project, documents and repository are fictional. In this demo the AI answers were simulated locally; the platform's engine, quote checks, scoring and reports are the real ones.
See which required work products exist for each framework and which are still missing.
Coverage, shared evidence and overlaps of all the project's frameworks, with frameworks that would fill the gaps.
Included frameworks come with curated rules and artifacts. For paid standards, you load your own licensed copy.
Rules are compared only within the same lifecycle stage and linked as equivalent, narrower, overlapping, complementary, preceding, producing or contradicting, with confidence and a written analysis.
Add keys for one or more providers, enable their models and assign a model to each task, with fallbacks. Each provider gets a maximum data classification.
Source reference, verbatim text, paraphrased structure and interpretations are labelled by origin, so reviewers know what came from the standard.
Re-running an assessment only reprocesses what changed.
Describe it, its criticality and its current lifecycle stage.
Included ones, your licensed copies, or your own policies turned into rules.
Upload documents and connect the repository; files are classified automatically.
Element-by-element verdicts with quotes, confidence and priority.
Recommendations, corrective actions and a compliance plan.
Apply proposed patches and draft missing documents from templates.
People confirm or override rules, relationships, drafts and findings.
Executive summary, results per framework and progress over time.
Openly licensed frameworks ship with curated rules. ISO/IEC 29110 ships as a rule extract; the standard's text is not redistributed.
For frameworks you buy or download, VirteVal guides you to the source and generates rules from the copy you upload.
And anything else
Internal policies, client requirements and contractual clauses can be turned into rules the same way.
Rules, evidence reading and first verdicts are automated; experts focus on review.
Quotes, documents and line ranges make results reproducible and auditable.
One assessment covers several frameworks, without repeating work.
Prioritised actions, corrections and templates move the project forward.
The team uploads its plans, requirements and test reports, connects its repository, and gets a gap list with templates for the missing work products.
A prioritised plan before the external assessor arrives.
Security leads run one assessment across three frameworks, with code findings and CI evidence from the repository.
One evidence base answers three frameworks.
Assessors share the same rules and criteria across projects, organised by organisation and team, and review AI verdicts before reporting.
Comparable results, each backed by evidence.
Project roles
Creates the project, chooses frameworks and manages the team.
Manages evidence, assessments and corrective actions.
Uploads documents and connects repositories.
Confirms or overrides findings, rules and drafts.
Reads results and reports.
AI providers
Development tools
General AI chat tools give opinions. Checklists give boxes to tick. VirteVal gives verdicts tied to rules and evidence.
Included, licensed or your own, assessed in one run with crosswalks between them.
No quote, no verdict. Insufficient evidence is reported as such.
Process artifacts and repositories are evaluated against the same rules.
| Need | Manual gap analysis | General AI chat | VirteVal |
|---|---|---|---|
| Rules from a standard | Hand-built checklist | Paraphrased from memory | Generated with verified source quotes and human approval |
| Evidence behind a verdict | Assessor's notes | Rarely cited | Literal quote linked to the document fragment |
| Several frameworks | Repeated per framework | One conversation at a time | One assessment with crosswalks and evidence reuse |
| Source code | Separate tools | Pasted snippets | Repository scans, CI evidence and SARIF |
| Choice of AI | Not applicable | One vendor | 12 provider presets, a model per task, your own keys |
Pricing depends on the number of projects and on whether you use your own AI keys. We quote after a demo with one of your projects.
No. It prepares you for certification or formal assessment by showing gaps and evidence. Certification is granted by accredited bodies.
No. ISO/IEC 29110 is included as a rule extract without the standard's text. For other paid standards you upload your own licensed copy.
Verdicts require a literal quote from your evidence, and missing evidence produces "insufficient evidence" rather than a guess. Reviewers can confirm or override every finding.
Yes. Use your own keys for any supported provider, set a data classification per provider, or point VirteVal at a self-hosted model through Ollama or any compatible API.
Yes. The interface, prompts and email templates are available in English and Spanish.
In the demo we assess a sample or your own documents and repository, and walk through verdicts, evidence, crosswalks and templates.