The LegalTek Trust Mark™
Verified Technology for a Trusted Legal Future
The LegalTek Trust Mark™ is the foundation of trust within LegalTek's ecosystem — a signal that a technology, firm, or workflow has met the highest standards for data integrity, ethical governance, and verified human oversight.
It tells clients, partners, and regulators one simple thing: LegalTek technology is tested, transparent, and accountable.
What the Trust Mark Means Inside LegalTek
Inside the LegalTek SaaS platform, the Trust Mark is not decorative. It's a live compliance engine and ranking signal embedded across every application, integration, and AI-powered workflow.
Each LegalTek-certified system carries a tiered verification badge:
LegalTek Verified™
Baseline data security and provenance assurance.
LegalTek Verified+™
Human-in-the-loop verification and continuous audit compliance.
LegalTek Assured™
Independently audited systems meeting the highest NIST and GDPR standards.
Every tier is earned, not claimed. The Trust Mark dynamically updates as audits, controls, and human-verification reviews are completed. It provides a real-time measure of compliance maturity.
What the Trust Mark Means Outside LegalTek
Externally, the Trust Mark serves as a recognition and ranking system for the broader LegalTek community — law firms, AI vendors, and data partners using LegalTek tools or participating in the LegalTek Marketplace.
When displayed on websites, reports, or SaaS integrations, the Trust Mark signals:
Roadmap: LegalTek intends to publish a public verification portal so clients and regulators can review the criteria a given badge represents. That verification portal is planned and is not yet live; today, each Trust Mark is issued by LegalTek based on the documented review below and is not an independent third-party audit.
The Framework Behind the Trust Mark
Every system evaluated for the LegalTek Trust Mark™ is assessed through two proprietary governance frameworks:
COUNSEL Framework
Confidentiality · Oversight · Understanding · Notification · Scrutiny · Equity · Lifetime Learning
COUNSEL ensures that every LegalTek process protects user privacy, provides clear notice of data use, undergoes human oversight, and promotes ongoing ethical improvement.
Explore the COUNSEL FrameworkG3M Framework
Govern · Map · Measure · Manage
G3M operationalizes governance — mapping data flows, measuring compliance, and managing ongoing risk across vendors and integrations.
Explore the G3M FrameworkTogether, they form the operational backbone of LegalTek's trust and compliance architecture.
Criteria Checked and How It's Verified
Data Provenance and Lineage
Every document, record, and AI output is cryptographically time-stamped and tracked through immutable logs. This chain of custody is reviewed quarterly to ensure that:
- Data origin and source identity are verifiable.
- Transformations (e.g., model input/output) are reproducible.
- No unverified data sources are introduced into LegalTek systems.
Human Verification Protocols
Before any client-facing AI result is released, licensed human reviewers validate outputs for:
- Legal sufficiency and contextual accuracy.
- Absence of bias, hallucination, or misrepresentation.
- Proper application of jurisdictional law and firm policy.
These reviews are sampled, logged, and used to train LegalTek's verification heuristics, which continuously improve oversight efficiency.
Audit and Logging Systems
All Trust-Marked entities maintain tamper-evident audit logs following NIST SP 800-92 and ISO 27035 incident reporting protocols. Logs record:
- Who accessed or modified data.
- When actions occurred.
- What system rules or models were active at the time.
Audits are performed by LegalTek's internal compliance team and, at the Assured tier, verified by independent third-party assessors.
Cybersecurity and Infrastructure Controls
LegalTek requires:
- End-to-end encryption (TLS 1.3+)
- Multi-factor authentication (MFA)
- Role-based access control
- Penetration testing under NIST 800-115 standards
- Roadmap: pursue annual third-party security review (e.g., SOC 2 Type II) for critical systems.
Failures or anomalies automatically suspend the Trust Mark until remediation is confirmed.
Vendor and Model Governance
Vendors and AI model providers integrated into LegalTek must sign Data Processing and Retention Agreements specifying:
- Model training boundaries (no unconsented client data).
- Retention and deletion timelines.
- Data residency jurisdiction (U.S. or GDPR-aligned regions).
- Certification of subcontractors under equivalent standards.
Data Retention and Destruction Policies
LegalTek's data lifecycle policies follow NIST SP 800-88 and GDPR Article 5(1)(e):
- Retention: typically seven years for evidentiary compliance.
- Deletion: irreversible cryptographic wiping after expiration.
- Exceptions: retained only under legal hold or regulatory mandate.
How the Trust Mark Is Earned
Application and Disclosure
The product or firm submits all relevant documentation on governance, privacy, and infrastructure.
COUNSEL Review
LegalTek's internal team evaluates privacy, oversight, and fairness alignment.
G3M Assessment
Technical auditors verify mapping, risk measurement, and process management.
Certification Review
LegalTek reviews controls and confirms the applicable compliance tier. Independent third-party audits are on the roadmap and not yet in operation.
Issuance and Monitoring
A digital Trust Mark is issued with a certification ID and a documented review record. A public verification portal and cryptographic time-stamping of issuance events are roadmap items, not current operational features.
Ongoing Monitoring
Any security incident, data-handling breach, or major product update triggers re-evaluation.
How the Trust Mark Affects Rankings
Within the LegalTek SaaS platform, the Trust Mark directly affects ranking visibility and user preference scoring:
Externally, LegalTek intends to offer a public verification interface — a planned, roadmap capability that would allow external firms, regulators, or clients to check a vendor's Trust Mark status. This public API is not yet live and is not a current operational endpoint. Today, verification is documented internally and handled directly by LegalTek on request; the future interface is designed to extend the Trust Mark's credibility into the wider AI and legal technology industry once released.
Connector Readiness: The Next Layer of AI Trust
A legal AI system is only as safe as the products, connectors, permissions, terms, data pathways, and human-review gates around it.
The LegalTek Trust Mark™ evaluates more than the model. It evaluates the operating environment around the model. In modern legal AI systems, that environment increasingly includes connectors: approved pathways between an AI assistant and a firm's email, calendar, document repositories, research tools, e-signature systems, meeting transcripts, project platforms, financial systems, and workflow automations.
For lawyers, connector readiness is not a convenience issue. It is a professional responsibility issue. A poorly governed connector can expose client information, blur privilege boundaries, create unauthorized workflows, preserve the wrong records, or allow an AI system to act without adequate human review.
LegalTek treats connectors as part of the Trust Mark analysis because connected AI is where theoretical risk becomes operational risk.
Connector governance also requires layered terms review. A product should not be evaluated only by its features. It should be evaluated by its standalone product terms, its AI-platform connector pathway, the permissions it grants, the records it creates, the retention rules it follows, the subprocessors it uses, and the human-review controls available to the firm.
Three Separate Review Layers
A product may pass one layer and fail another. LegalTek evaluates each layer independently before recommending legal AI use.
Standalone Product Terms
The vendor's own terms, privacy policy, DPA, MSA, AI terms, retention rules, subprocessor list, and customer-data commitments.
Can the firm use this product for legal work at all?
Claude Connector Review
The Claude connector pathway, Anthropic terms, MCP rules, workspace controls, permissions, tool access, and action authority.
Can Claude safely access this product for the intended legal workflow?
ChatGPT Apps / Connectors / Actions Review
The ChatGPT app or connector pathway, OpenAI terms, admin controls, actions, connected-app permissions, custom app rules, and audit controls.
Can ChatGPT safely access this product for the intended legal workflow?
Why Connectors Matter
Access
What systems can the AI reach? LegalTek evaluates whether access is limited by user role, matter, workspace, and business purpose.
Data Flow
What data can move into or out of the system? LegalTek reviews source systems, retention rules, training boundaries, and whether client-identifying information is protected.
Action Authority
Can the AI only retrieve information, or can it also draft, send, file, schedule, delete, label, update, or trigger automations? Higher-risk actions require stronger approval gates.
Terms Control
What terms govern each layer? LegalTek reviews standalone product terms, AI-platform terms, connector rules, data-processing commitments, subprocessors, security, retention, audit logs, confidentiality, and enterprise controls.
Connector Readiness Crosswalk
Connector governance fits directly within the LegalTek Trust Mark™ because every integration changes the firm's risk profile.
| Trust Mark Dimension | Connector Readiness Question | Legal Risk Controlled |
|---|---|---|
| Confidentiality | Does the connector expose client-identifying facts, privileged communications, or confidential work product? | Confidentiality and privilege protection |
| Oversight | Can a lawyer review and approve outputs before client-facing, court-facing, or system-changing use? | Supervision, unauthorized practice, and filing risk |
| Understanding | Does the firm understand what the connector can access, retrieve, store, and trigger? | Technology competence and vendor due diligence |
| Notice | Should the client be told that a connected AI system is being used in the matter? | Client communication and informed consent |
| Scrutiny | Are connected outputs verified against source documents and primary authority? | Hallucination, citation, and misstatement risk |
| Equity | Are automation savings billed ethically and transparently? | Fee reasonableness and billing integrity |
| Learning | Is connector performance, risk, and vendor behavior periodically re-reviewed? | Ongoing competence and policy maintenance |
| Terms Review | Do the governing terms permit the intended legal use case and protect client data? | Vendor contracting, confidentiality, privilege, and data-use risk |
| Action Control | Can the AI take external actions, or is it limited to retrieval and drafting? | Unauthorized action, deletion, filing, communication, and workflow-trigger risk |
| Auditability | Can the firm prove what happened, who approved it, and what sources were used? | Recordkeeping, supervision, dispute response, and defensibility |
AI Platform Terms References
Standalone product terms are only one part of the review. When an AI platform accesses a product through an app, connector, custom connector, MCP server, or action, the AI platform's own terms and controls must also be reviewed.
ChatGPT / OpenAI
Directory availability is not Trust Mark approval. A product's standalone terms, Claude connector availability, and ChatGPT app, connector, or action availability must be reviewed separately. LegalTek treats each layer as a distinct governance question involving vendor terms, AI platform terms, data handling, retention, permissions, auditability, action authority, and human oversight. Terms and legal links are provided for review only and may change without notice.
Last reviewed: July 9, 2026. Product terms, connector availability, app availability, platform policies, and provider terms may change. LegalTek.ai is not affiliated with Anthropic, Claude, OpenAI, ChatGPT, or any listed connector provider.
Standalone Product Terms Review
This grouped list identifies the standalone vendor terms and legal references that should be reviewed before any product is approved for legal AI workflows. These links do not mean the vendor is approved. They are starting points for LegalTek's Trust Mark review.
Last reviewed: July 9, 2026. Product terms, platform policies, connector availability, app availability, scopes, permissions, and admin controls may change. LegalTek recommends periodic re-review before enabling or expanding connected AI workflows.
Default Legal AI Connector Posture
For legal workflows, LegalTek generally recommends starting with the least authority necessary. Default to read-only access where possible. Disable sending, signing, filing, deleting, paying, publishing, or triggering external automations unless the workflow has been expressly reviewed. Require human approval before client-facing, court-facing, financial, or system-changing action.
Admin Controls Determine Whether a Connector Is Defensible
Even acceptable product terms may not be enough. LegalTek also reviews whether the firm can control who enables the connector, what scopes are granted, which workspaces or folders are accessible, whether write actions can be disabled, whether activity is logged, and whether administrators can revoke access.
Custom Connectors Require Heightened Review
Custom connectors, MCP servers, custom apps, and actions may connect AI systems to tools that are not part of a public directory or have not been independently verified by the AI platform. LegalTek treats custom connectors as higher-risk unless the firm can verify the endpoint, authentication model, data flow, logging, retention, vendor terms, and action authority.
Terms Evidence LegalTek Collects Per Product
The public page currently shows the primary Terms / Legal link. A full Trust Mark review record captures the following evidence, when available, and is maintained privately per product.
Recommended Law Firm Starter Stack
For most small and mid-size law firms, connector readiness does not mean connecting everything. It means connecting the right systems first, under the right terms, with the right permissions and review gates.
LegalTek generally recommends starting with:
- Microsoft 365 or Google Workspace
- Docusign
- One legal research system
- One governed document repository
- One meeting-intelligence workflow
- One internal collaboration platform
- One controlled automation platform
The goal is not maximum connectivity. The goal is defensible connectivity.
Before a Connector Can Be Trusted
Is Your AI Stack Connector-Ready?
LegalTek can evaluate your firm's AI tools, standalone product terms, Claude connector pathways, ChatGPT Apps / Connectors / Actions, vendor terms, permissions, and workflow controls against the LegalTek Trust Mark™ rubric. The result is a practical green/yellow/red roadmap for safer legal AI adoption.
Agentic AI: From Chatbots to Autonomous Workflows
Agentic AI systems do more than answer questions. They can plan, use tools, call APIs, retrieve records, remember context, trigger workflows, and sometimes act across connected systems.
Traditional AI tools generate responses. Agentic AI systems pursue goals. They can decide what steps to take, what tools to use, what information to retrieve, and when to ask a human for approval. That shift creates enormous opportunity for law firms, but it also creates new duties around confidentiality, supervision, permissions, auditability, vendor terms, and human oversight.
LegalTek evaluates agentic AI because a law firm\u2019s risk is no longer limited to what the model says. The risk includes what the agent can see, what it can remember, what it can change, what it can send, what it can trigger, and whether the lawyer remains in control.
Agentic AI is the next shift in legal technology. Instead of simply responding to prompts, agentic systems can pursue goals, break work into steps, choose tools, retrieve information, call APIs, update systems, remember context, and coordinate with other agents. A chatbot may draft an answer. An agent may search a file system, check a calendar, summarize a transcript, draft an email, prepare a task list, update a CRM, trigger an automation, or ask another specialized agent to complete part of the work.
LegalTek treats agentic AI as a Trust Mark issue because agentic systems move AI from “answer generation” into “workflow execution.”
Connector readiness determines what an AI system can access. Agentic AI readiness determines what the system can do with that access.
Three-Layer Governance Stack
Connectors are the access layer. Agents are the action layer. The Trust Mark is the governance layer.
What Makes AI Agentic?
1. Goal
An agent is given an objective, not just a one-time question. Example: “Prepare a client intake summary and flag missing information.”
2. Planning
The agent breaks the objective into steps: review intake form, search emails, check calendar, identify documents, draft follow-up questions.
3. Tool Use
The agent calls tools, apps, APIs, databases, calendars, document systems, or web services to complete each step.
4. Memory and State
The agent may retain context across steps, sessions, matters, users, or systems — including drafting style, project history, and prior results.
5. Action
The agent may create, update, send, file, publish, delete, trigger, or route work — draft an email, schedule a meeting, or prepare a filing packet.
The governance risk increases as the agent moves from \u201Cread\u201D to \u201Cwrite,\u201D from \u201Cdraft\u201D to \u201Csend,\u201D and from \u201Csuggest\u201D to \u201Cact.\u201D
Agentic AI Maturity Ladder
| Level | Name | Description | LegalTek Governance Posture |
|---|---|---|---|
| Level 0 | Chatbot | Responds to prompts but does not use tools or act externally. | Standard AI output review required. |
| Level 1 | Tool-Assisted Assistant | Uses approved tools to retrieve or analyze information. | Verify sources, scopes, and access permissions. |
| Level 2 | Workflow Agent | Executes multi-step workflows across one or more systems. | Require logs, role controls, and workflow review. |
| Level 3 | Multi-Agent System | Coordinates specialized agents for research, drafting, review, or operations. | Require orchestration controls and human checkpoints. |
| Level 4 | Action Agent | Can write, send, update, delete, publish, sign, pay, or trigger external actions. | Require explicit approval gates and action restrictions. |
| Level 5 | Autonomous Operating Agent | Runs persistently, monitors triggers, and acts over time with limited prompting. | High-risk; enterprise controls and documented governance required. |
LegalTek generally recommends that law firms begin with Level 1 or Level 2 agentic workflows and avoid Level 4 or Level 5 authority unless the use case has been formally reviewed.
Agentic AI Landscape: Examples LegalTek Tracks
The agentic AI market includes personal agents, self-hosted agents, multi-agent frameworks, workflow-orchestration platforms, enterprise governance platforms, and business-app agent systems. LegalTek tracks these categories to help firms distinguish useful automation from uncontrolled autonomy.
These examples are not approvals. They are market examples for governance review. Inclusion does not mean LegalTek approval, Trust Mark approval, legal suitability, security approval, or client-data approval. Agent capabilities, product terms, platform terms, privacy commitments, security controls, and availability may change. LegalTek evaluates agentic systems case by case.
Common Agentic Design Patterns
Single-Agent Workflow
One agent receives a goal, uses tools, and completes a task.
Best for: Low-risk research, summarization, document extraction, or intake triage.
Risk: The agent may overstep if tools are too broad.
Multi-Agent Crew
Multiple specialized agents collaborate. One may research, another drafts, another reviews, another verifies.
Best for: Complex research, drafting, review, and operations workflows.
Risk: Errors can compound if agents trust each other without source verification.
Human-in-the-Loop Agent
The agent pauses before high-risk actions and waits for lawyer approval.
Best for: Client communication, court-facing work, filing, contract approval, payment, or publication.
Risk: The approval step must be real, logged, and easy to understand.
Event-Triggered Agent
The agent runs when something happens — a new email, form submission, calendar event, signed agreement, or CRM update.
Best for: Intake, follow-ups, task routing, CRM updates, and status tracking.
Risk: Bad triggers can create unauthorized or repetitive actions.
Persistent Agent
The agent runs over time, monitors systems, remembers context, and may act across sessions.
Best for: Operations monitoring, research watchlists, matter-status support, and executive assistants.
Risk: Memory, retention, credential use, and ongoing autonomy require heightened governance.
Self-Improving Agent
The agent updates skills, prompts, tool descriptions, or workflows based on experience.
Best for: Experimental internal tools and advanced technical teams.
Risk: Behavior drift, unapproved changes, regression failures, and auditability problems.
Agentic AI Risk Matrix
| Risk Area | Why It Matters | LegalTek Review Question |
|---|---|---|
| Data Access | Agents may retrieve confidential, privileged, or client-identifying data. | What data sources can the agent access, and are they scoped by role, matter, or workspace? |
| Tool Permissions | Agents may call tools that read, write, send, delete, publish, sign, pay, or trigger workflows. | What tools can the agent use, and which actions require human approval? |
| Memory | Agents may retain context across sessions or users. | What does the agent remember, where is it stored, and how can it be deleted? |
| Planning | Agents may choose their own path to complete a goal. | Are plans constrained by policy, playbooks, or approved workflows? |
| Multi-Agent Delegation | Agents may rely on other agents’ work. | Are intermediate steps and sources visible to the lawyer? |
| External Actions | Agents may affect real systems. | Can the agent send, file, share, publish, delete, sign, pay, or trigger automations? |
| Auditability | The firm may need to prove what happened. | Are prompts, tool calls, sources, outputs, approvals, and timestamps logged? |
| Vendor Terms | Product and platform terms may govern data use differently. | Have standalone product terms and AI-platform terms been reviewed separately? |
| Human Oversight | Lawyers remain responsible for legal work. | Where is human review mandatory, and is it documented? |
| Unauthorized Practice | Agents may appear to provide legal advice. | Are disclaimers, supervision, and client-facing boundaries in place? |
| Billing Ethics | Automation may change time, value, and billing practices. | Are fees reasonable, transparent, and consistent with professional duties? |
| Incident Response | Agents can make mistakes across multiple systems quickly. | Can the agent be stopped, audited, rolled back, and contained? |
Data Access
Agents may retrieve confidential, privileged, or client-identifying data.
Ask: What data sources can the agent access, and are they scoped by role, matter, or workspace?
Tool Permissions
Agents may call tools that read, write, send, delete, publish, sign, pay, or trigger workflows.
Ask: What tools can the agent use, and which actions require human approval?
Memory
Agents may retain context across sessions or users.
Ask: What does the agent remember, where is it stored, and how can it be deleted?
Planning
Agents may choose their own path to complete a goal.
Ask: Are plans constrained by policy, playbooks, or approved workflows?
Multi-Agent Delegation
Agents may rely on other agents’ work.
Ask: Are intermediate steps and sources visible to the lawyer?
External Actions
Agents may affect real systems.
Ask: Can the agent send, file, share, publish, delete, sign, pay, or trigger automations?
Auditability
The firm may need to prove what happened.
Ask: Are prompts, tool calls, sources, outputs, approvals, and timestamps logged?
Vendor Terms
Product and platform terms may govern data use differently.
Ask: Have standalone product terms and AI-platform terms been reviewed separately?
Human Oversight
Lawyers remain responsible for legal work.
Ask: Where is human review mandatory, and is it documented?
Unauthorized Practice
Agents may appear to provide legal advice.
Ask: Are disclaimers, supervision, and client-facing boundaries in place?
Billing Ethics
Automation may change time, value, and billing practices.
Ask: Are fees reasonable, transparent, and consistent with professional duties?
Incident Response
Agents can make mistakes across multiple systems quickly.
Ask: Can the agent be stopped, audited, rolled back, and contained?
Agent Action Authority Scale
The more authority an agent has, the stronger the governance requirement.
| Authority Level | Agent Can | LegalTek Default Posture |
|---|---|---|
| Read | Search, retrieve, summarize, classify | Generally lowest risk, but still requires access controls. |
| Draft | Prepare proposed text, tasks, emails, memos, forms | Human review required before use. |
| Write | Create or update internal records | Restrict to approved workflows with logs. |
| Send / Share | Send emails, share files, message users, publish updates | Human approval required. |
| Delete / Modify | Delete files, overwrite records, change settings | Disable unless specifically reviewed. |
| Sign / Pay / File | Execute signatures, payments, filings, or legal submissions | Critical risk; require formal approval gates and documented review. |
| Trigger Automation | Launch cross-system workflows | High to critical risk depending on workflow; require scoped triggers and audit logs. |
Read
Search, retrieve, summarize, classify
Default: Generally lowest risk, but still requires access controls.
Draft
Prepare proposed text, tasks, emails, memos, forms
Default: Human review required before use.
Write
Create or update internal records
Default: Restrict to approved workflows with logs.
Send / Share
Send emails, share files, message users, publish updates
Default: Human approval required.
Delete / Modify
Delete files, overwrite records, change settings
Default: Disable unless specifically reviewed.
Sign / Pay / File
Execute signatures, payments, filings, or legal submissions
Default: Critical risk; require formal approval gates and documented review.
Trigger Automation
Launch cross-system workflows
Default: High to critical risk depending on workflow; require scoped triggers and audit logs.
Human-in-the-Loop Workflow
Approved governance path
No direct client-facing, court-facing, financial, filing, publishing, signing, deleting, or system-changing action without approval.
Agentic AI Risk Heat Map
The highest-risk workflows combine sensitive legal data with external action authority.
| Data Sensitivity ↓ / Authority → | Read-only | Draft | External Action |
|---|---|---|---|
| Privileged / Financial / Litigation | Medium | High | Critical |
| Confidential | Medium | High | High |
| Internal | Low | Medium | High |
| Public | Low | Low | Medium |
Default LegalTek Agentic AI Posture
For legal workflows, LegalTek generally recommends starting with the least authority necessary. Begin with read-only or draft-only access. Disable sending, signing, filing, deleting, paying, publishing, or triggering external automations unless the workflow has been expressly reviewed. Require human approval before client-facing, court-facing, financial, or system-changing action.
The goal is not maximum autonomy. The goal is supervised, auditable, defensible autonomy.
Before an Agentic AI System Can Be Trusted
Every agentic workflow reviewed under the LegalTek Trust Mark™ must clear the questions below. Human-approval gates and auditability items are non-negotiable for any authority beyond read or draft.
Scope & Purpose
Define what the agent is for before granting any authority.
- What is the agent’s narrowly-defined purpose?
- What legal workflows is it approved for — and explicitly not approved for?
- What systems, matters, folders, channels, projects, or workspaces can it access?
- What data can it retrieve, and what data is off-limits (privileged, sealed, PII, PHI, client-confidential)?
- What tools, APIs, and connectors is it allowed to call?
Memory & Learning
Constrain what the agent remembers, learns, or changes about itself.
- Does it have persistent memory, and is that memory scoped per user, matter, or client?
- Can memory be inspected, edited, exported, and purged on request?
- Can it learn from prior sessions, fine-tune, or modify its own skills or prompts?
- Are self-modifications gated by human review before they take effect?
Human-Approval Gates
No client-facing, court-facing, financial, or system-changing action without a human decision.
- Is human approval required before sending any external communication (email, message, portal post)?
- Is human approval required before signing, filing, or serving any document?
- Is human approval required before payment, transfer, refund, or any financial action?
- Is human approval required before publishing, posting, or sharing externally?
- Is human approval required before deleting, overwriting, or modifying records of legal significance?
- Is human approval required before triggering downstream automations, workflows, or other agents?
- Is client consent or notice captured for the workflow, model, and data flows involved?
- Is the reviewer clearly identified (name, role, timestamp) at each gate?
- Is there a documented “no-approval, no-action” default for anything above read or draft?
Auditability & Evidence
Every action must be reconstructable after the fact.
- Are prompts, system instructions, retrieved sources, tool calls, and outputs logged with timestamps?
- Are approvals, denials, edits, and overrides logged with the human reviewer’s identity?
- Are model, version, temperature, and configuration recorded for each run?
- Are logs tamper-evident, access-controlled, and retained per firm and client obligations?
- Can a full action trail be exported for a given matter, client, user, or date range?
- Are outputs verified against source documents and primary authority before use?
- Are hallucinations, refusals, and safety interventions tracked as first-class events?
Containment & Response
Assume something will go wrong — be able to stop, contain, and recover.
- Can the agent be disabled globally, per user, per matter, or per tool within minutes?
- Is there a documented rollback plan for actions the agent has already taken?
- Is there an incident-response runbook covering wrong-recipient sends, wrong-account posts, unauthorized filings, and data exposure?
- Are clients, opposing counsel, courts, and regulators notified when required?
- Is the workflow re-reviewed whenever terms, tools, models, permissions, or subprocessors change?
Terms, Data & Governance
The paperwork must match the action authority.
- Are standalone product terms reviewed against LegalTek Trust Mark™ criteria?
- Are AI-platform terms reviewed separately from product terms?
- Are data processing, retention, deletion, and subprocessor terms reviewed and mapped to the workflow?
- Is training-on-customer-data disabled or contractually restricted where required?
- Is the agent’s authority scale (Read / Draft / Send / Sign / Pay / File / Trigger) documented and reviewed?
Default rule: if a checklist item cannot be answered with evidence, the agent stays at read or draft authority. No sending, signing, filing, paying, publishing, deleting, or triggering downstream automations until every human-approval gate and auditability item is verifiably in place.
The public Trust Mark page explains LegalTek\u2019s agentic AI review framework. Product-specific determinations, contract analysis, security findings, workflow diagrams, and final approval status should be documented in a private Trust Mark review report, not inferred from a public market list.
Last reviewed: July 9, 2026. Agentic AI products, platform capabilities, vendor terms, connector availability, admin controls, and security practices may change.
Is Your Firm Ready for Agentic AI?
LegalTek can evaluate your firm\u2019s AI agents, automation workflows, connected systems, vendor terms, permissions, memory, audit logs, and human-review controls against the LegalTek Trust Mark\u2122 rubric.
Vendor Terms Evaluation
LegalTek evaluates third-party AI tools against the COUNSEL Framework before recommending them for legal workflows. Below is our assessment of productivity and legal-research vendors commonly considered by attorneys.
Analysis based on publicly posted Terms of Service and Privacy documentation. This is not legal advice; firms must conduct independent due diligence before deploying any vendor with client data.
Strengths
- • Layered Terms structure (Platform, Application, User) separates enterprise from consumer obligations.
- • API documentation is public and permits programmatic access with issued keys.
- • Distinguishes personal use from organization-bound acceptance for authorized signatories.
Concerns for Legal Use
- • Meeting audio and AI summaries constitute attorney-client communications when used on privileged calls — retention and access disclosures must be scrutinized before enabling on client matters.
- • Terms do not publicly commit to no-training on customer meeting content at the free/personal tier; enterprise DPAs must be negotiated separately.
- • Consent-to-record obligations shift entirely to the user; Ohio and most one-party-consent states are covered, but all-party-consent jurisdictions (CA, FL, IL, PA, WA) require affirmative disclosure.
COUNSEL Crosswalk
LegalTek Assessment: Acceptable for internal, non-privileged meetings. Not currently eligible for the LegalTek Trust Mark™ for use on client-privileged calls without a negotiated enterprise agreement, BAA (where PHI is possible), and documented all-party consent workflows.
Sources & Audit Trail — Granola2 documents · click to expand
| Document | Exact URL | Publisher date | Retrieved by LegalTek |
|---|---|---|---|
| Application Terms of Servicemonitored | https://docs.granola.ai/help-center/policies/terms-of-service/application-terms-of-service | not stated | Aug 6, 2026 |
| Privacy Policy | https://docs.granola.ai/help-center/policies/privacy-policy | not stated | Aug 6, 2026 |
- Evaluation record
- v1.0 · —
- Last automated URL check
- no check recorded yet
- Check result
- —
Retrieval dates record when LegalTek read the document text that this assessment relies on. Publishers may revise terms without notice; the automated checker verifies only that the URL still resolves, not that the text is unchanged. Verify the live document before relying on this assessment. Not legal advice.
Wispr Flow
Voice-to-text dictation · Evaluation v1.0 · Last updated —
Strengths
- • Explicit user ownership of Inputs and assignment of Outputs back to the user (§2.B).
- • Third-party LLM providers (OpenAI, Anthropic) contractually barred from training on user data; 30-day deletion (§6).
- • User-controllable model-training toggle via published Data Controls Policy.
- • Privacy Mode: no dictation stored on Wispr servers.
- • SOC 2 Type II, ISO 27001, and HIPAA certified per public privacy disclosures.
- • Prohibits AI Features for automated legally-significant decisions without adequate human review (§3.B) — aligned with COUNSEL Oversight.
- • Prohibits recording third parties without consent (§3.A).
Concerns for Legal Use
- • Standard "AS IS" warranty disclaimer on Output accuracy (§12) — attorneys remain solely responsible for verifying dictated content before filing.
- • Broad license to "reproduce, modify, distribute, transmit, export, display, store" Customer Content for service operation (§2.B). Confirm scope in enterprise agreement for privileged dictation.
- • Third-Party Outputs clause disclaims uniqueness — dictation refined by generative AI may resemble other users' Outputs (§2.D).
- • Limitation of liability excludes consequential damages (§13) — standard SaaS but relevant for malpractice-exposure analysis.
COUNSEL Crosswalk
LegalTek Assessment: Materially stronger data-governance posture than most productivity vendors. Provisionally aligned with LegalTek Verified™ baseline for attorney dictation workflows when Privacy Mode is enabled and model-training is disabled at the account level. Firms should still execute a BAA before dictating any PHI and confirm data-residency requirements for regulated matters.
Sources & Audit Trail — Wispr Flow2 documents · click to expand
| Document | Exact URL | Publisher date | Retrieved by LegalTek |
|---|---|---|---|
| Terms of Servicemonitored | https://wisprflow.ai/terms-of-service | not stated | Aug 6, 2026 |
| Privacy Policy | https://wisprflow.ai/privacy | not stated | Aug 6, 2026 |
- Evaluation record
- v1.0 · —
- Last automated URL check
- no check recorded yet
- Check result
- —
Retrieval dates record when LegalTek read the document text that this assessment relies on. Publishers may revise terms without notice; the automated checker verifies only that the URL still resolves, not that the text is unchanged. Verify the live document before relying on this assessment. Not legal advice.
Midpage
AI legal research & drafting · Evaluation v1.0 · Last updated —
Strengths
- • Explicit public commitment: Midpage does not use customer data to train or fine-tune AI models.
- • Agreements with AI model providers contractually bar those vendors from training on Midpage customer data.
- • SOC 2 Type II certified with annual independent validation.
- • AES-256 encryption at rest and TLS 1.2+ in transit.
- • Plugin integrations do not store submitted queries, uploads, or outputs.
- • Source-grounded RAG architecture generates answers exclusively from reviewed cases with hyperlinked, verifiable citations — materially reducing Rule 1.1 hallucination risk.
Concerns for Legal Use
- • As a venture-stage vendor, enterprise DPA and BAA must be confirmed in writing before processing client-confidential or PHI matters.
- • Model providers (subprocessors) may retain submitted queries for up to 60 days per disclosed subprocessor terms.
- • Web app stores product data including user queries, uploaded materials, and generated outputs while the account remains active; deletion is within 60 days of account closure or valid request.
- • Attorneys must independently verify citations and analysis before filing — the platform is a research accelerator, not a substitute for professional judgment.
COUNSEL Crosswalk
LegalTek Assessment: Strong data-governance posture for a venture-stage legal-AI vendor. The explicit no-training commitment, SOC 2 Type II certification, and source-grounded RAG architecture make it a promising candidate for LegalTek Verified™ status once a signed DPA and BAA (where applicable) are confirmed. Firms should enable plugin integrations where possible to minimize data retention.
Sources & Audit Trail — Midpage2 documents · click to expand
| Document | Exact URL | Publisher date | Retrieved by LegalTek |
|---|---|---|---|
| Terms and Conditionsmonitored | https://www.midpage.ai/terms-and-conditions | not stated | Aug 6, 2026 |
| Privacy Policy | https://www.midpage.ai/privacy-policy | not stated | Aug 6, 2026 |
- Evaluation record
- v1.0 · —
- Last automated URL check
- no check recorded yet
- Check result
- —
Retrieval dates record when LegalTek read the document text that this assessment relies on. Publishers may revise terms without notice; the automated checker verifies only that the URL still resolves, not that the text is unchanged. Verify the live document before relying on this assessment. Not legal advice.
CourtListener
Free legal research database · Evaluation v1.0 · Last updated —
Strengths
- • Non-profit mission (Free Law Project) with no commercial incentive to monetize user data.
- • Core service hosts only public court opinions — no client-confidential data is ingested into an AI training pipeline through ordinary search use.
- • Does not sell visitor information; usage logs are deleted automatically at 12 weeks.
- • Transparent removal policy: documents are not removed without a court order, protecting the integrity of the public record.
- • Anonymized data may be shared with academic researchers for system improvement, but not for commercial AI model training.
Concerns for Legal Use
- • Terms explicitly disclaim reliability and accuracy of reproduced documents — attorneys must independently verify every citation and quotation before filing.
- • "AS IS" warranty disclaimer with no representations regarding truthfulness, accuracy, or relevancy of communicated information.
- • Broad limitation of liability excludes direct, indirect, incidental, special, and consequential damages.
- • Service may be discontinued or altered without prior notice; no guarantee of perpetual availability.
- • Searches and usage patterns are logged (albeit anonymized after 12 weeks) — consider this for sensitive research topics.
COUNSEL Crosswalk
LegalTek Assessment: CourtListener is a public-interest research tool, not a confidential data processor. For researching public court opinions, it presents no Rule 1.6 confidentiality risk. The primary COUNSEL duty implicated is Understanding — attorneys must independently verify citations because the Terms explicitly disclaim accuracy. Recommended for general legal research and citation verification workflows; not a substitute for verified primary-source review before filing.
Sources & Audit Trail — CourtListener1 document · click to expand
| Document | Exact URL | Publisher date | Retrieved by LegalTek |
|---|---|---|---|
| Terms of Service & Privacy Policy (combined)monitored | https://www.courtlistener.com/terms/ | not stated | Aug 6, 2026 |
- Evaluation record
- v1.0 · —
- Last automated URL check
- no check recorded yet
- Check result
- —
Retrieval dates record when LegalTek read the document text that this assessment relies on. Publishers may revise terms without notice; the automated checker verifies only that the URL still resolves, not that the text is unchanged. Verify the live document before relying on this assessment. Not legal advice.
Even Realities
Smart glasses / wearable HUD hardware · Evaluation v1.0 · Last updated —
Strengths
- • Terms are publicly hosted on the vendor's support portal and clearly identify governing law, dispute resolution, and account-termination rights.
- • Hardware-first form factor (prescription-ready smart glasses with on-lens HUD) means most day-to-day use does not require continuous cloud capture.
- • Consumer purchase model — no automatic corporate telemetry ingestion of firm systems.
- • Companion-app permissions are user-controllable at the OS level (Bluetooth, mic, notifications).
Concerns for Legal Use
- • Terms are consumer-focused: broad "AS IS / AS AVAILABLE" warranty disclaimer and limitation of liability. No enterprise DPA or BAA is publicly offered.
- • No public no-training commitment covering audio, imagery, or conversational data captured via the glasses or companion app.
- • Wearable capture in privileged settings triggers Rule 1.6 confidentiality duties and, in all-party-consent jurisdictions (CA, FL, IL, PA, WA), affirmative disclosure to every participant.
- • Most federal and state courthouses prohibit recording devices in courtrooms and secure areas — the glasses can qualify as a recording device even when passive.
- • Third-party AI services integrated through the companion app may be governed by separate terms outside Even Realities' control.
COUNSEL Crosswalk
LegalTek Assessment: Acceptable as a personal HUD for non-privileged reading, navigation, and translation. Not recommended for wear during client meetings, depositions, mediations, or courtroom appearances without (1) a negotiated enterprise agreement addressing training and retention, (2) firm-level device and consent policies, and (3) affirmative all-party disclosure in every applicable jurisdiction. Not currently eligible for the LegalTek Trust Mark™ for privileged workflows.
Sources & Audit Trail — Even Realities1 document · click to expand
| Document | Exact URL | Publisher date | Retrieved by LegalTek |
|---|---|---|---|
| Terms of Service (support article)monitored | https://support.evenrealities.com/hc/en-us/articles/14290554040335-Terms-of-Service | not stated | Aug 6, 2026 |
- Evaluation record
- v1.0 · —
- Last automated URL check
- no check recorded yet
- Check result
- —
Retrieval dates record when LegalTek read the document text that this assessment relies on. Publishers may revise terms without notice; the automated checker verifies only that the URL still resolves, not that the text is unchanged. Verify the live document before relying on this assessment. Not legal advice.
Honcho
AI memory & personalization API (Plastic Labs) · Evaluation v1.0 · Last updated —
Strengths
- • Unusually transparent privacy documentation: named subprocessor table (Supabase, Fly.io, Stripe, Vercel, Sentry, PostHog, AWS) including optional inference providers (Anthropic, Groq, Google Cloud).
- • Explicit opt-in-only training posture: public model training on customer content does not occur without affirmative opt-in.
- • Stated encryption controls: TLS 1.2+ in transit and AES-256 at rest.
- • Published retention schedule with configurable content retention (90-day default), immediate hard-delete purge endpoint, and 90-day backup expiry.
- • Open-source (AGPL-3.0) build available for self-hosting, which keeps memory data inside firm-controlled infrastructure.
Concerns for Legal Use
- • Vendor states plainly that the hosted service is not certified under SOC 2, HIPAA, or comparable frameworks — a material gap for firms with client security addenda.
- • Non-public fine-tuning on de-identified data is contemplated; de-identification of legal matter context is difficult to assure.
- • The product's purpose is to persist user context over time, so privileged facts accumulate by design rather than incidentally.
- • Hosting is primarily US-based; data-residency commitments for cross-border matters are not enumerated.
- • No formal DPA offering is stated in the public Terms; must be confirmed in writing.
COUNSEL Crosswalk
LegalTek Assessment: Best-in-class disclosure hygiene for an early-stage infrastructure vendor, but the absence of any independent certification keeps the hosted tier outside Trust Mark eligibility for privileged data. Firms building matter-aware assistants should prefer the self-hosted AGPL build inside firm infrastructure, or restrict the hosted API to non-privileged, non-client context.
Sources & Audit Trail — Honcho2 documents · click to expand
| Document | Exact URL | Publisher date | Retrieved by LegalTek |
|---|---|---|---|
| Terms of Servicemonitored | https://app.honcho.dev/tos | not stated | Aug 6, 2026 |
| Privacy Policy | https://app.honcho.dev/privacy | not stated | Aug 6, 2026 |
- Evaluation record
- v1.0 · —
- Last automated URL check
- no check recorded yet
- Check result
- —
Retrieval dates record when LegalTek read the document text that this assessment relies on. Publishers may revise terms without notice; the automated checker verifies only that the URL still resolves, not that the text is unchanged. Verify the live document before relying on this assessment. Not legal advice.
Buzz (Block, Inc.)
Nostr-based human + AI-agent workspace, developer preview · Evaluation v1.0 · Last updated —
Strengths
- • Bring-your-own-model architecture: Block states plainly that it does not provide, resell, or host any AI model, so no Block-side inference layer touches matter content by default.
- • The content license granted to Block is limited to operating, securing, moderating, and improving the Service — there is no separate grant to train foundation models on customer content.
- • The license terminates when content is deleted, subject only to legally required preservation, limited-duration backups, and copies others already made.
- • Customer retains ownership of all content; the Terms expressly disclaim any transfer of ownership.
- • The Buzz client is open source and self-hostable, and the Terms confirm self-hosted instances fall outside the Block-operated Service entirely — a viable firm-infrastructure deployment path.
- • The agent provisions are unusually explicit about human accountability: no deceptive presentation of agents as humans, and express allocation of oversight duty to the operator.
Concerns for Legal Use
- • Developer preview. Features are offered on a preview, beta, or experimental basis, provided "as is," and may be changed or removed — a poor fit for matter-critical systems.
- • Relay architecture is the central Rule 1.6 issue: content published to a Buzz Relay is stored, indexed, and distributed under the Nostr protocol, and the Service may connect to third-party relays Block neither controls nor includes in the Service.
- • Agent-to-model egress is uncontrolled by the vendor: when an agent runs, content it can access — including other users' messages in any channel that agent can reach — may be transmitted to the model provider the customer chose, under that provider's terms, not Block's.
- • No key recovery. Identity is a cryptographic keypair; Block states it generally cannot restore a lost or compromised key, so key loss can mean permanent loss of matter records.
- • Anyone holding a private key can act as the holder — an authentication model that demands firm-managed key custody policy before any client data enters a workspace.
- • Admins bear the compliance burden: access control, permissions, monitoring, and enforcement inside a workspace are contractually the Admin's responsibility.
- • Mandatory individual arbitration and a class-action waiver (opt-out available), plus broad warranty disclaimers and liability limits.
- • Feedback assignment is total — all feedback becomes Block's exclusive property, including any IP rights in it.
- • No SOC 2, ISO 27001, HIPAA, or BAA availability appears in the public Terms, and no DPA is referenced for the hosted tier.
COUNSEL Crosswalk
LegalTek Assessment: Architecturally interesting and, on the training question, better drafted than most agent platforms — Block takes no training license and hosts no model. The disqualifier for privileged data is structural, not editorial: a protocol-based relay that indexes and distributes events, an explicit possibility of third-party relays outside the Service, and agent-to-model egress that can carry other users' channel content to whichever provider the customer connected. Under Rule 1.6 and Rule 5.3, that is three trust boundaries the firm must govern and only one the vendor controls. Not eligible for the LegalTek Trust Mark™ for client-confidential work in the hosted developer preview. Firms interested in the model should evaluate the self-hostable open-source client inside firm infrastructure, with a written key-custody policy, a channel-scoping policy limiting what any agent can read, and a connected-model provider that itself carries a no-training commitment.
Sources & Audit Trail — Buzz (Block, Inc.)2 documents · click to expand
| Document | Exact URL | Publisher date | Retrieved by LegalTek |
|---|---|---|---|
| Terms of Servicemonitored | https://buzz.xyz/terms-of-service | Jul 17, 2026 (stated) | Aug 6, 2026 |
| Privacy Policy | https://buzz.xyz/privacy-policy | not stated | Aug 6, 2026 |
- Evaluation record
- v1.0 · —
- Last automated URL check
- no check recorded yet
- Check result
- —
Retrieval dates record when LegalTek read the document text that this assessment relies on. Publishers may revise terms without notice; the automated checker verifies only that the URL still resolves, not that the text is unchanged. Verify the live document before relying on this assessment. Not legal advice.
Obsidian
Local-first Markdown knowledge base (optional Sync/Publish) · Evaluation v1.0 · Last updated —
Strengths
- • Local-first architecture: notes are plain Markdown files on the attorney's own device, so ordinary use involves no vendor-side processing of matter content.
- • No first-party AI model trains on note content, because the core application does not transmit vault content to the vendor.
- • Optional cloud add-ons only: Sync and Publish are separately purchased and separately scoped.
- • Deletion terms for the paid tiers are specific — data removed immediately on user cancellation, or one month after a lapsed subscription.
- • Plain-file storage supports firm records-retention and matter-closure workflows without vendor dependency.
Concerns for Legal Use
- • Legal documents are thin and templated — the Privacy Policy still refers to the developers' earlier product, Dynalist, rather than Obsidian.
- • Security language is generic; no TLS version, at-rest algorithm, or contractual end-to-end-encryption commitment appears in the legal text. Sync's end-to-end encryption is documented in help and marketing pages, not in the Terms.
- • Subprocessors are not named; the policy relies on a general assurance that third parties are bound by appropriate agreements.
- • The real AI exposure is the community plugin ecosystem: third-party AI plugins can transmit vault content to outside model providers under their terms, entirely outside Obsidian's control.
- • No DPA, BAA, or security certification is offered in public documentation.
COUNSEL Crosswalk
LegalTek Assessment: Recommended for attorney work-product capture in a local-only configuration on an encrypted, firm-managed device. Sync and Publish, and any AI plugin, should be treated as separate vendor decisions requiring their own review — Publish in particular renders notes publicly, an obvious Rule 1.6 hazard if a matter vault is misconfigured. Firms should maintain an approved-plugin allowlist before permitting AI extensions on matter vaults.
Sources & Audit Trail — Obsidian2 documents · click to expand
| Document | Exact URL | Publisher date | Retrieved by LegalTek |
|---|---|---|---|
| Terms of Servicemonitored | https://obsidian.md/terms | not stated | Aug 6, 2026 |
| Privacy Policy | https://obsidian.md/privacy | not stated | Aug 6, 2026 |
- Evaluation record
- v1.0 · —
- Last automated URL check
- no check recorded yet
- Check result
- —
Retrieval dates record when LegalTek read the document text that this assessment relies on. Publishers may revise terms without notice; the automated checker verifies only that the URL still resolves, not that the text is unchanged. Verify the live document before relying on this assessment. Not legal advice.
Agent OS
Open-source spec framework for coding agents (Builder Methods) · Evaluation v1.0 · Last updated —
Strengths
- • Not a hosted service: Agent OS is a set of Markdown standards and spec files installed into the user's own repository, so there is no vendor-side collection, retention, or training pipeline.
- • MIT-licensed open source, permitting firm-internal modification and audit of every instruction file.
- • Runs entirely inside the developer's environment and chosen coding agent.
- • Its purpose — forcing explicit specifications and standards before an agent acts — is structurally aligned with COUNSEL Oversight and documented-process discipline.
Concerns for Legal Use
- • No Terms of Service and no Privacy Policy are published; the only governing instrument located is the MIT license on the public repository. Firms requiring a vendor contract have nothing to countersign.
- • MIT terms disclaim all warranties and all liability — the entire risk of agent behavior sits with the firm.
- • The real data exposure belongs to the underlying coding agent (for example Claude Code or Cursor), which transmits repository context under its own terms; Agent OS neither constrains nor discloses that flow.
- • Open-source instruction files are mutable supply-chain inputs — updates should be pinned and reviewed rather than pulled blindly.
- • Nothing in the framework enforces confidentiality; placing client data in a repo governed by Agent OS specs does not protect it.
COUNSEL Crosswalk
LegalTek Assessment: Evaluated as a development methodology rather than a data processor. Safe to adopt for firm-side engineering because it introduces no new vendor data flow, but it confers no confidentiality protection and carries no vendor accountability. Trust Mark review should be directed at the underlying coding agent and at the firm's repository controls, not at Agent OS itself.
Sources & Audit Trail — Agent OS2 documents · click to expand
| Document | Exact URL | Publisher date | Retrieved by LegalTek |
|---|---|---|---|
| MIT License (governing terms)monitored | https://github.com/buildermethods/agent-os/blob/main/LICENSE | not stated | Aug 6, 2026 |
| Project repository & documentation | https://github.com/buildermethods/agent-os | not stated | Aug 6, 2026 |
- Evaluation record
- v1.0 · —
- Last automated URL check
- no check recorded yet
- Check result
- —
Retrieval dates record when LegalTek read the document text that this assessment relies on. Publishers may revise terms without notice; the automated checker verifies only that the URL still resolves, not that the text is unchanged. Verify the live document before relying on this assessment. Not legal advice.
Side-by-Side Summary
| Criterion | Granola | Wispr Flow | Midpage | CourtListener | Even Realities | Honcho | Buzz (Block) | Obsidian | Agent OS |
|---|---|---|---|---|---|---|---|---|---|
| User owns Output | Not explicit in public terms | Yes — assigned to user | Not explicit in public terms | N/A (public data) | Not addressed (consumer hardware) | Customer content licensed to vendor for service delivery | Yes — ownership expressly retained | Yes — files stay on device | Yes — MIT license, user's repo |
| No training on customer data | Enterprise-tier only | User-toggle + third-party barred | Explicit no-training commitment | N/A (no AI training pipeline) | Not publicly committed | Opt-in only for public models | No training license to Block | N/A (no first-party AI) | N/A (not a service) |
| Public security certifications | Not publicly enumerated | SOC 2 Type II · ISO 27001 · HIPAA | SOC 2 Type II | Non-profit / 501(c)(3) | Not publicly enumerated | None — vendor states not SOC 2 / HIPAA | None found; developer preview | None found | N/A (open-source framework) |
| Local/on-device privacy mode | Cloud-processed | Privacy Mode available | Plugin mode stores nothing | N/A (web search) | On-device HUD; cloud services via companion app | Self-hostable (AGPL-3.0) | Self-hostable open-source client | Local-first by default | Fully local files |
| Human-review requirement for legal AI use | Not addressed | Contractually required | Attorney-directed by design | Required (disclaimed reliability) | Required (broad AS-IS disclaimer) | Not addressed (developer API) | Operator oversight duty stated in Terms | Attorney-authored notes | Spec-first review encouraged |
| Preliminary Trust Mark eligibility | Not recommended for privileged workflows without enterprise DPA | Provisionally LegalTek Verified™ with Privacy Mode | Promising candidate once DPA confirmed | Recommended for public-record research | Not recommended for privileged use | Self-hosted only for privileged context | Not eligible in hosted preview; self-host for privileged context | Recommended in local-only configuration | Out of scope — review the underlying coding agent |
The Promise of Verified Trust
The LegalTek Trust Mark™ embodies the company's mission:
"To build technology that empowers lawyers and protects clients — through governance, not guesswork."
Every mark issued is a badge of discipline, transparency, and earned trust.
Inside or outside LegalTek, it means one thing above all: verified integrity in the age of intelligent law.