Product Vision | Intactus Docs

What Is Intactus?

Intactus is an AI-native operating system for trial attorneys at small and mid-size law firms. It automates the capture, preservation, search, analysis, and court-ready packaging of digital evidence, with tamper-evident chain of custody built in at the point of capture.

The name is Latin for "untouched, unaltered, intact." The evidence enters the system and remains exactly as captured (hashed, timestamped, and manifested) while the platform makes it searchable, analyzable, and ready for court.

The Problem

For Attorneys

Attorneys handling cases with digital evidence (text messages, emails, social media posts, blog content, audio, video) face a fragmented, manual, and error-prone process:

Capture is manual. Attorneys and clients take screenshots one at a time. Content gets deleted before it can be preserved. There is no systematic, ongoing monitoring.
Chain of custody is an afterthought. Screenshots saved to a desktop folder have no forensic integrity. Authenticity can be challenged in court.
Search and organization is painful. Hundreds of pieces of evidence across texts, emails, screenshots, and documents. Attorneys dig through folders manually.
Analysis is slow and expensive. Identifying patterns, timelines, and contradictions across a large evidence corpus requires hours of attorney time or expensive expert witnesses.
Court preparation is tedious. Assembling exhibit packages, building timelines, and generating reports for motions and hearings is manual labor.
AI tools risk privilege. After the Heppner ruling, consumer AI tools waive attorney-client privilege. There is no privilege-safe AI evidence tool built for small firms.
Evidence work is practice overhead. Every hour spent on evidence administration (intake, organization, search, timeline building, exhibit preparation) is an hour not spent on legal reasoning, advocacy, or taking on more cases.

For Clients

The people who need legal help face an equally broken process, and it's the reason many never take action:

The documentation burden is overwhelming. 400 text messages, 50 emails, months of social media posts. The legal system asks people to read, screenshot, organize, and present all of it, on top of an already stressful situation.
Documentation is retraumatizing. Every screenshot means re-reading content that is threatening, manipulative, or emotionally damaging. People stop documenting, or stop pursuing legal action entirely, because the process itself is an ordeal.
Evidence is fragmented across platforms. Texts on the phone, emails in Gmail, posts on Instagram, blogs on Substack, voicemails in the phone app. No single tool captures across all of them.
Clients are in the dark about their own cases. Every question costs $300/hour. Clients hold back, feel anxious, and make uninformed decisions because staying informed feels too expensive.
Consumer AI tools create liability. After Heppner, any case research a client does through ChatGPT or Claude is non-privileged and potentially discoverable.

This problem exists across practice areas: family law, protective orders, harassment, defamation, and any case where digital evidence matters.

What Is Intactus?

The Problem

For Attorneys

Attorneys handling cases with digital evidence (text messages, emails, social media posts, blog content, audio, video) face a fragmented, manual, and error-prone process:

Capture is manual. Attorneys and clients take screenshots one at a time. Content gets deleted before it can be preserved. There is no systematic, ongoing monitoring.
Chain of custody is an afterthought. Screenshots saved to a desktop folder have no forensic integrity. Authenticity can be challenged in court.
Search and organization is painful. Hundreds of pieces of evidence across texts, emails, screenshots, and documents. Attorneys dig through folders manually.
Analysis is slow and expensive. Identifying patterns, timelines, and contradictions across a large evidence corpus requires hours of attorney time or expensive expert witnesses.
Court preparation is tedious. Assembling exhibit packages, building timelines, and generating reports for motions and hearings is manual labor.
AI tools risk privilege. After the Heppner ruling, consumer AI tools waive attorney-client privilege. There is no privilege-safe AI evidence tool built for small firms.
Evidence work is practice overhead. Every hour spent on evidence administration (intake, organization, search, timeline building, exhibit preparation) is an hour not spent on legal reasoning, advocacy, or taking on more cases.

For Clients

The people who need legal help face an equally broken process, and it's the reason many never take action:

The documentation burden is overwhelming. 400 text messages, 50 emails, months of social media posts. The legal system asks people to read, screenshot, organize, and present all of it, on top of an already stressful situation.
Documentation is retraumatizing. Every screenshot means re-reading content that is threatening, manipulative, or emotionally damaging. People stop documenting, or stop pursuing legal action entirely, because the process itself is an ordeal.
Evidence is fragmented across platforms. Texts on the phone, emails in Gmail, posts on Instagram, blogs on Substack, voicemails in the phone app. No single tool captures across all of them.
Clients are in the dark about their own cases. Every question costs $300/hour. Clients hold back, feel anxious, and make uninformed decisions because staying informed feels too expensive.
Consumer AI tools create liability. After Heppner, any case research a client does through ChatGPT or Claude is non-privileged and potentially discoverable.

This problem exists across practice areas: family law, protective orders, harassment, defamation, and any case where digital evidence matters.

The Solution

Intactus provides an end-to-end pipeline with two integrated surfaces: a client portal for evidence intake and case understanding, and an attorney dashboard for evidence management, analysis, and court preparation.

1. Client Portal (Mobile-First)

The client's primary interface. Designed to minimize evidence capture friction: the client submits everything and the system handles the rest.

Multi-modal evidence capture: Text message exports (iMessage, Android, WhatsApp), email forwarding to a case-specific intake address, URL submission for forensic web capture, file upload for screenshots/documents/audio/video
Incident log: Quick-entry form for documenting new events ("he was 40 minutes late for pickup again") without writing a detailed narrative
Chat with My Case: Ask factual questions about evidence ("how many times was he late for pickup?") and get answers grounded in the actual evidence vault, without incurring attorney fees. Queries run through enterprise infrastructure with zero data retention (ZDR), so client interactions are not exposed to third-party consumer AI terms
Secure messaging: Encrypted, asynchronous communication with the attorney and staff, linked to the case record
Case visibility: Timeline view, evidence upload status, access to exhibits and reports the attorney has shared
Privacy by design: Magic link authentication, evidence stored off-device

2. Evidence Vault (Attorney-Facing)

Every artifact is hashed (SHA-256), timestamped, and recorded in a tamper-evident JSONL manifest
Chain of custody documented at the point of capture
Tagged by case, source, type, and relevance
Full-text search across all evidence in a case
Pre-computed entity extraction (names, dates, locations, dollar amounts)
Pre-computed relationship extraction (who sent what to whom, who mentioned whom)
Pre-computed classification (threatening, financial, custody-relevant, etc.)
Auto-generated 2-line summaries of each piece of evidence

3. Attorney-Directed Analysis (AI Layer)

Attorney initiates all analysis, maintaining attorney-client privilege and work product doctrine
Agentic retrieval: The LLM reasons over a structured evidence index to find relevant items. No vector database, no embedding vendor, full provenance on every result.
Synthesis: "Summarize the pattern across these 15 emails"
Timeline: "Build a chronological narrative from these events for my motion"
Comparison: "Compare this witness's social media posts against their deposition testimony"
Detection: "Identify financial discrepancies between these posts and the sworn disclosure"
Practice-area analysis lenses selected by the attorney
All AI runs through Anthropic's commercial API with contractual confidentiality protections. Prompts and completions are not used for model training under commercial API terms. Migration to AWS Bedrock with PrivateLink and zero data retention is planned before onboarding paying customers beyond the design partner.

4. Court-Ready Output

Evidence packages with SHA-256 hash verification instructions
Chronological timelines with source citations
Reports suitable for motions, hearings, and trial
PDF export with chain-of-custody documentation
Exhibit preparation with Bates-style numbering (Future)

5. AI Agents as Practitioners

AI agents work alongside the attorney's team, reviewing evidence, drafting reports, and answering client questions, all under attorney direction and all audit-logged. Unlike competitors that treat AI as a feature layer (a chatbot here, a summarization button there), Intactus treats agents as first-class participants with the same operational vocabulary as human users.

Intake Agent triages new uploads, extracts metadata, and flags items for review
Research Agent searches evidence, finds patterns, identifies contradictions
Drafting Agent generates court-ready reports by composing atomic tools, not running hardcoded pipelines
Client-Facing Agent powers "Chat with My Case" within the client portal
Every agent action is logged with attorney attribution, model identity, and reasoning trace
Agents are shaped by layered instruction profiles: firm defaults, attorney preferences, case-type presets, and case-specific context merge automatically. A Research Agent working a custody case for Attorney A applies different analytical priorities than the same agent type working a defamation case for Attorney B — without either attorney writing a single prompt.
Attorney directs, agents execute, and the audit trail proves it. Heppner-safe by design.

6. Practice Efficiency

Beyond individual cases, Intactus changes how the attorney's practice operates:

Eliminate evidence administration: Automated intake, organization, search, and timeline generation replace 5-10 hours of manual work per case
Handle more cases without hiring: The platform replaces paralegal hours on evidence work with software that costs a fraction of a salary
Reduce client management overhead: The client portal answers factual questions, provides case visibility, and handles secure communication, so the attorney's time goes to legal judgment, not information retrieval
Accelerate client intake: Send the prospective client a portal link during the first call. By the meeting, evidence is ingested, organized, and searchable. Case assessment in minutes, not weeks.
AI that learns your practice: Instruction profiles encode the attorney's analytical framework, severity scales, output preferences, and practice-area knowledge. Every AI operation — from evidence classification to report generation to client Q&A — automatically applies the attorney's judgment. The profiles compound: the more the attorney uses the platform, the more precisely the AI reflects their standards. Onboarding a new associate means installing the firm's profile, not six months of mentorship.

Target Market

Primary ICP: Small and mid-size law firms (1-20 attorneys) practicing:

Family law: divorce, custody, child support, paternity
Protective orders / harassment: TROs, injunctions, stalking, domestic abuse
Defamation / online libel: content preservation before deletion

Design partner: A boutique litigation and trial firm in Madison, WI practicing family law.

Secondary markets (expansion):

Criminal defense (witness impeachment, discovery review)
Employment law (hostile work environment, discrimination)
Legal aid organizations serving DV/harassment victims (grant-funded)
Personal injury firms
IP / trademark attorneys
General civil litigation

Why Now?

The Heppner ruling (Feb. 10, 2026): SDNY ruled that documents created with consumer AI tools are not privileged. This creates urgent demand for enterprise-grade, privilege-safe AI legal tools. Intactus's architecture is designed to be on the right side of this ruling.
DocuSAFE shut down (Oct. 2023): The most widely recommended free DV evidence app ceased operations with no replacement. The victim self-service market is unserved.
Legal tech spending surged 9.7% in 2025: fastest growth likely ever in the legal industry.
AI adoption gap: 72% of solo attorneys use AI in some capacity, but only 8% have adopted it deeply. Purpose-built tools that "just work" for a specific use case can close this gap.
No integrated pipeline exists: Attorneys currently stitch together 3-5 separate tools (capture, storage, search, analysis, export) to do what Intactus does in one platform.
Text messages are the new evidence frontier: Texts are the #1 evidence type in family law and harassment cases, but no tool makes them easy to capture, preserve, and analyze at scale.

What Makes This Defensible

Chain of custody lock-in: Once evidence packages and hashes are in the system, court submissions reference Intactus's chain-of-custody documentation. Migrating mid-case is unthinkable.
Manifest and hash chain: Proprietary to the platform's format. Artifacts can be independently verified, but the system of record is Intactus.
Practice-area depth: Specialized analysis lenses (harassment pattern detection, financial contradiction flagging, custody evidence organization) are hard for general-purpose tools to replicate.
Privilege-by-design architecture: Designed to maintain attorney-client privilege at every layer: single AI vendor (Anthropic), zero data retention, PrivateLink, attorney-directed analysis. The Heppner-safe design is a structural advantage over tools that bolt on legal compliance as an afterthought.
Agent-native architecture: The platform is designed so that every human action has an agent equivalent and every agent action has an audit trail. Competitors treat AI as a feature layer: a chatbot on a document viewer, a summarization button on a search screen. Intactus's tool layer, agent permission model, and audit infrastructure are structural properties that are hard to retrofit, giving it a first-mover advantage in agent-native legal tech.
No embedding vendor dependency: Agentic retrieval over structured evidence indexes eliminates the need for a second AI provider. One contract, one ZDR agreement, one privilege argument.
Personalized AI that compounds: Instruction profiles encode each attorney's analytical framework, preferred formats, risk tolerances, and practice-specific patterns into layered instruction files that customize every AI operation. A firm's accumulated judgment becomes a durable, transferable asset — new hires produce work product calibrated to the firm's standards from day one. General-purpose AI tools require the attorney to re-explain their preferences every session; Intactus remembers and applies them automatically. The deeper an attorney invests in their profiles, the harder it is to leave.
Client portal as moat: The client experience, including frictionless capture, Chat with My Case, and secure messaging, creates value that makes the attorney's firm stickier. The client doesn't want to switch attorneys because the evidence platform is too useful to leave.

Origin

Intactus grew out of a real forensic evidence archival project: a system built to capture, preserve, and analyze digital evidence of cyberstalking and harassment for use in legal proceedings. That system automated Substack blog archival, Gmail forensic capture, audio/video interception, SHA-256 tamper-evident hashing, and AI-powered threat analysis, producing law-enforcement-ready evidence packages.

The attorney retained for that case saw the system and immediately asked about purchasing it as a service. Intactus is the productization of that working prototype.