DeepJudge vs Fileread: how they compare in 2026

There is no content asset and no conventional product underneath. Unlike the other research records in this lane, DeepJudge licenses nothing and publishes nothing of its own: the material it works on belongs entirely to the customer and stays in the customer's systems. What DeepJudge supplies is the retrieval itself, described as intent-based search that understands content, context and relevance, with automatic classification into a taxonomy and near-duplicate and redline detection across collections the vendor describes as billions of unstructured and structured data points. Strip the models out and what remains is a set of connectors and a permissions synchronizer, which is plumbing rather than something a firm would buy. The founders' own framing, that search rather than models is the wedge, is a claim about which machine learning matters, not a claim that machine learning is peripheral. Verified 2 September 2026.

The product is the model work and nothing survives its removal. Fileread exists to read a document production and answer questions about it: users query the full collection in plain language rather than by keyword, and the vendor's own framing is that the point is asking the question you cannot turn into a keyword. Generation is the deliverable, not a feature bolted onto a repository: fact memos, chronologies, contradiction reports, character lists, org charts, deposition outlines and factual briefs are produced from the record, and a module called Workbench runs longer chains of research to assemble them. Retrieval spans formats that defeat keyword tooling, including images, handwritten notes, audio, video and spreadsheets, which is model work rather than indexing. The vendor's positioning sentence is an AI claim about its own architecture, that most legal AI is built to generate and this one is built to be checked, with every answer pinned to the location in the source. Strip the models out and there is no residual product: no case law database, no docket service, no document management system, only a customer's own production which the customer already holds. Homepage, product framing and the vendor's human oversight paper read 7 September 2026.

Grounding is structural rather than asserted: every result is a document the firm already holds, in its own system, which the user can open and read in full, so the verification path is the shortest of any record in this pull. The retrieval method is described at a working level, covering intent-based search, automatic taxonomy classification, near-duplicate and redline detection, and multi-level combination of client, matter, document and people data. Section 4(j) of the subscription terms places a contractual obligation on the customer to verify, including through human review, the accuracy, reliability and appropriateness of the AI features. What is absent is measurement. No accuracy figure, evaluation, test set or benchmark appears anywhere, and nothing addresses hallucination in the workflow and agent layer, which is where generated text rather than retrieved documents is produced. The one published number nearby is a satisfaction measure, that 90 percent of users report finding results faster than with existing tools, which is not an accuracy claim. Verified 2 September 2026.

Grounding is real, documented and architecturally central, and no accuracy figure exists. What is published: every answer returns pinned to the exact location in the source document, a claim traces to its page so checking takes a glance, and generated work product retains its connection to the underlying material so a reviewer can inspect where a finding came from. The vendor states the design intent plainly, that the system is built to be checked rather than to be trusted, and maintains a dedicated paper on source citation and verification. It also publishes an authored, dated and practitioner-reviewed taxonomy of seven ways legal AI fails, with a detection control for several of them: unsupported generation, correct source with incorrect interpretation, context loss, missed evidence, conflicting evidence presented as certainty, identity confusion, and date and sequence errors. That is an unusually candid document and one of the failure modes has a product answer, since contradiction reports exist to surface the conflicts that the fifth mode describes being flattened. What keeps this off the top grade is the distinction between naming the failure modes of a class and measuring your own. The taxonomy is written as guidance about legal AI generally, not as disclosure about this system: no accuracy figure, no error rate, no test set, no benchmark and no evaluation result is published for Fileread itself. A buyer can see that the vendor understands how such systems fail and cannot see how often this one does. Homepage, failure modes paper of 17 August 2026 and human oversight paper read in full 7 September 2026.

The human review requirement is contractual and the excluded uses are unusually explicit. Section 4(j) requires the customer to verify the accuracy, reliability or appropriateness of AI features through human review, and the same clause prohibits use in judicial decision-making by judges, in the professional activities of lawyers serving as arbitrators, mediators or other alternative dispute resolution neutrals, and in any use restricted under the EU Artificial Intelligence Act. Naming the roles that may not use the product is rarer and more useful than the usual disclaimer. What is not published is the operating boundary. The workflow layer is described as agentic reasoning that executes complex multi-step tasks, retrieving from multiple sources and taking actions, and the customer builds those workflows itself using a low-code builder, so the threshold at which a workflow acts without a human is set by the firm and described nowhere. Governance of agents is offered as a capability without any account of what the governance controls are. Verified 2 September 2026.

Substantive published commitments with the autonomy boundary stated, short of anything the product enforces. The boundary is stated, which is what keeps this off the middle grade: Workbench performs longer chains of research and assembles chronologies, factual briefs, contradiction reports and deposition outlines, and the vendor states that the resulting work remains a draft for professional review and refinement. So the reader is told the system runs multi-step work unattended and is told where it stops. Around that sits a published oversight framework, authored, dated and practitioner-reviewed, which is more than most records on this axis offer: oversight is framed as beginning before the prompt, with the task, included and excluded materials, intended audience, acceptable level of uncertainty, claims requiring source support, the person responsible for final review and the record to be retained all settled in advance; review is scaled to risk, with exploratory output to be treated as a starting point, citations sampled, and results not represented as established fact; and a twelve-point review checklist runs from whether cited passages actually support the statements to whether a responsible professional approved the final use. What is absent is the product half. The framework is guidance offered to the buyer rather than a specification of the system's own controls, only one paragraph describes what Fileread itself does, no configurable autonomy level or override is described, and nothing states that the checklist is recorded or enforced anywhere in the product. Human oversight paper and homepage read in full 7 September 2026.

Sixteen firms are named on the home page, spanning global elite practices including Freshfields, Greenberg Traurig, Holland and Knight and Gunderson Dettmer, United States mid-market firms including Cozen O'Connor, ArentFox Schiff and Vorys, Swiss and Austrian practices including Homburger, Lenz and Staehelin, Wenger Vieli, Advestra, Schoenherr and CMS Switzerland, and one in-house function at SBB. Three named individuals speak on the record with title and firm, including two chief innovation officers. Figures are published: 65 hours saved per user per year on searching, 85 percent adoption within two months, 23 queries per user per day, fourfold return in the first year, and 90 percent reporting faster results. One figure is properly attached, with a named M&A partner at Homburger stating that more than 80 percent of the firm's legal professionals had incorporated it into their workflow. What holds this below A is that the headline figures are aggregate and carry no method or measurement basis, and no dates appear on the home page; the individual customer story pages were not opened this pass. Verified 2 September 2026.

Named customers with named people in named roles, and one quantified account, with no method behind any of it. Five attributed testimonials are published: a partner at Kellogg, Hansen, Todd, Figel & Frederick describing use on an expedited case that went to trial; a partner at Fields, Han & Cunniff describing the system retrieving a correct document from an inaccurate description; a senior vice president for innovation and strategy at JND Legal Administration assessing the product on maturity, defensibility and user experience against alternatives; a manager of knowledge and practice support services at Nutter McClennen & Fish on the Relativity integration; and the chief technology officer of Cimplifi on replacing Boolean search. That is a spread across a litigation boutique, a large firm practice-support function, a litigation-support provider and an eDiscovery services company, which tells a buyer who actually runs it. One quantified account is published with the arithmetic shown: a matter arrived as 4,000 text message screenshots, which the platform converted and deduplicated to 725 actual messages, work the vendor states would have consumed more than 20 hours of attorney time. Independent coverage is cited, including a Law360 piece on a firm's generative AI experiments and reporting of a 6 million dollar seed round. What is absent is measurement: no aggregate figures, no recall or precision on retrieval, no customer count, and the time saved rests on a single anecdote rather than a study. Homepage, customer testimonials and cited coverage read 7 September 2026.

Four of the five limbs are met, and the segregation limb is met better here than by any other record in this pull. Section 5(b) of the subscription terms provides that DeepJudge does not use customer data to train the AI or machine learning models underlying the services, which is a contractual prohibition naming the thing rather than a policy statement. Segregation is the product's architecture rather than a claim about it: access permissions are continuously synchronized with the firm's source systems so that existing ethical walls are maintained, which is the matter-level requirement a firm buyer needs, and section 5(e) adds that customer authored workflow logic will not be disclosed or incorporated into the service provided to any other customer. Retention and deletion are stated, with customer data deleted within sixty days of termination. The position on model providers is answered structurally, since the firm selects the models and the hybrid deployment clause discloses exactly when processing leaves the firm's own infrastructure. The limb that fails is privilege and work product handling: neither term appears anywhere, and ethical walls address conflicts rather than privilege. Verified 2 September 2026.

Real controls are published and nothing addresses privilege, which matters more here than on most records because the material this product ingests is an entire litigation production. What exists: a statement that nothing a customer uploads trains a shared model; data encrypted and hosted in Azure and processed through what the vendor calls private large language models; and a trust center control inventory naming least privilege, access monitoring, access log management and data asset classification, alongside HIPAA compliance which bears on the sensitive material productions contain. Those are corporate and infrastructure controls, and they are credited on the stewardship axis rather than counted twice here. What is absent is everything this axis asks. No privilege or work product treatment is published at any level. Nothing states who at the vendor may read customer matter material or under what conditions. No segregation position is published between matters held by the same customer or between customers, and the question is sharpened by the Relativity integration, since nothing states whether the platform inherits the workspace permissions a firm has already built there or maintains a second permission model of its own. The customer agreement that would carry a confidentiality clause is not readable: a Master Services Agreement is listed in the trust center behind an access request, and the website terms render client-side and returned no body on this channel. The vendor's own oversight paper names protecting privileged material as an attorney duty, which is an acknowledgment of the risk rather than a commitment about the product. Homepage, security page and trust center read 7 September 2026.

This is the most complete professional responsibility position located in the pull, and all of it sits in the operative agreement. Section 4(j) states that DeepJudge is not a law firm and does not engage in the practice of law and that the customer acknowledges it is not relying on the services for legal advice. It states who may use the product and, unusually, who may not, prohibiting use in judicial decision-making or other activities performed by judges, and in the professional activities of lawyers serving as arbitrators, mediators or other alternative dispute resolution neutrals. It reaches competence and supervision through a positive obligation rather than a disclaimer, requiring the customer to verify the accuracy, reliability or appropriateness of the AI features including through human review. And the legal boundary is stated twice: section 4(a) confines use to compliance with the obligations applicable where the customer and its users are located, and section 4(j) excludes any use prohibited or restricted under the EU Artificial Intelligence Act, naming the instrument. The coverage limb that asks about jurisdictional reach of legal content does not bite on a product that searches the firm's own documents rather than any jurisdiction's law, and is recorded here as inapplicable rather than failed. No bar or law society guidance is engaged, which is recorded on the signal rather than here. Verified 2 September 2026.

Substantive published engagement with where professional responsibility sits, short of a located disclaimer or a named authority. Several limbs of this band do not bite and are named rather than penalized: the product analyses the customer's own document production rather than advising on law, its outputs are factual findings cited to the record rather than legal conclusions citing authority, and it is sold to lawyers and litigation support professionals rather than to the public, so the unauthorised practice risk this axis was written for is structurally remote. On what does bite, the vendor engages directly and in its own authored material. Its oversight paper states that artificial intelligence can assist with parts of litigation work but does not possess responsibility for the representation; that attorneys must interpret contested facts, assess credibility, understand procedural posture, protect privileged material and apply legal standards; that a fluent response may obscure uncertainty, missing context or an incorrect assumption; and that an exploratory question for internal orientation does not require the same process as a factual representation intended for a client, a witness examination or a court filing. Its review checklist ends with the output being reviewed under applicable legal and professional standards and a responsible professional approving the final use. What is missing is the formal half: no disclaimer language was located, because the terms of service render client-side and returned no body, and no bar rule, ethics opinion or court standing order is named anywhere. Human oversight paper and failure modes paper read in full 7 September 2026.

One binding governance commitment is published and everything around it is missing. Section 4(j) of the subscription terms makes compliance with the EU Artificial Intelligence Act a contractual restriction on use, naming the instrument and excluding the administration-of-justice applications the Act treats as high risk, which is more than a principles page and is enforceable by the customer. The workflow layer is also sold partly on governance, offering to build, deploy, orchestrate and govern AI agents. What no published material supplies is any content behind either. No responsible AI page exists, no governance framework, no individual or function named as accountable for model behavior, no account of what is tested before a release ships, no certification such as ISO 42001, and nothing at all about uneven output across matter types, practices or populations. ISO 27001 and SOC 2 Type II are security attestations, which this axis treats as a different subject. Checked the home page, product page, security page, subscription terms and privacy policy on 2 September 2026. Verified 2 September 2026.

An AI-specific governance apparatus is disclosed by name, and its contents are gated. The trust center lists two instruments in its policy library that most records in this corpus do not have at all: an AI System Development and Evaluation Policy, and an Artificial Intelligence Management System Plan, the latter using the term of art for the management system that ISO/IEC 42001 defines. It carries a separate AI section with overview, security and monitoring entries, so AI is treated as its own control domain rather than folded into general security, and the whole trust center is described as continuously monitored with automated compliance monitoring and automated alert response. That is real substance, and it is the reason this is not the middle grade. What keeps it off the top is that nothing behind the titles was read and nothing is attested. Both policies sit behind an access request, so their existence and titles are established and their contents are not, and crediting a document by its title is not available here. No ISO/IEC 42001 certification is claimed, in contrast to ISO/IEC 27001:2022 which is claimed with its version. No evaluation result, benchmark or model card is published. Nobody is named as accountable for the system's behavior. And no bias, fairness or representativeness disclosure exists at any level, which is worth naming for a product whose identity confusion failure mode, published by the vendor itself, involves combining information belonging to people with similar names, addresses or roles. Trust center and failure modes paper read 7 September 2026.

All five elements are established and specific, on the subscription terms and the data processing addendum, version 1.1 of May 2026. Retention and deletion: section 3(a) of the subscription terms deletes all customer data including training data and customer models within sixty calendar days of termination, and DPA clause 11 adds prompt destruction or return on request, extending to data held by subprocessors, with confirmation of destruction. Access control: permissions continuously synchronized from source systems on least privilege principles, access restricted to authenticated users, configurable audit logging, indexing controls over what is crawled at all, and DPA Annex 2 adding multi factor authentication, role based access control, centralized event logging and segregated pre production and production environments. Subprocessors: Annex 3 names five entities with purpose and location, and clause 5 requires equivalent written terms, due diligence, vendor liability for their acts, thirty days notice before a new subprocessor processes anything, and an objection right with termination without liability. Incident practice: clause 7 commits to notice without undue delay and in any event within seventy two hours of discovery, covering breaches at subprocessors as well as at DeepJudge, with the definition at 1.13 expressly including unauthorised internal access, plus cooperation so the customer can meet its own regulator and data subject duties. Encryption is TLS 1.2 or later in transit and AES-256 at rest, with annual penetration testing and controls audited for SOC 2 Type II. Clause 3.4 adds a purpose limitation binding subprocessors: no selling or sharing, no processing for other purposes, no retention or disclosure outside the direct business relationship, and no combining customer personal data with data from other sources.

The published control inventory is the broadest located in this pull, and one limb is missing. Certifications are named with the standard version: SOC 2 Type 2, ISO/IEC 27001:2022, HIPAA, GDPR and CCPA. Training is addressed: nothing a customer uploads trains a shared model, with data encrypted and hosted in Azure and processed through private large language models. Subprocessors are named on the public page rather than on request, being OpenAI, Baseten, Pinecone, Langfuse and Microsoft Azure, with Amazon Web Services and Azure listed as infrastructure, and a subscribable feed publishes changes to that list with dated entries recording the addition of Langfuse for prompt use and of OpenAI. Incident practice is present, with a documented incident reporting process and designated response personnel. Access control is present, with least privilege, access monitoring, access log management and data asset classification. Around those sit data loss prevention, application and network penetration testing, code analysis, credential management, backups with a backup policy, business continuity and disaster recovery including contingency plan testing and failover, change management with change notification and verification, endpoint controls including mobile device management and threat detection, anti-DDoS and firewalls, and security, privacy and phishing training. The gap, named because every partial names its limitation, is retention and deletion: no retention period, no deletion commitment and no end-of-matter or end-of-engagement disposal position was located anywhere, which is the question a litigation buyer asks about a production it has uploaded. Trust center, security page and homepage read 7 September 2026.

The whole allocation of loss is published, specific, and mutual, which is rare in this corpus. Section 7(a)(ii) caps each party's total aggregate liability at the fees paid to DeepJudge in the twelve months before the event giving rise to it, and names four express exceptions that sit outside the cap: the customer's payment obligations, breach of the use restrictions at section 4(e), fraud or wilful misconduct, and infringement of intellectual property rights. Section 7(a)(i) excludes indirect and consequential damages symmetrically for both parties, with the same carve-outs. Section 7(b) gives the customer a real indemnity running the right way, with DeepJudge defending third-party claims that the services or underlying software infringe intellectual property rights, and setting out the remedies in order: procure the right to continue, modify or replace to become non-infringing, or terminate the affected portion. There is also a service commitment a buyer can hold: section 1(a)(iii) sets a 99 percent quarterly uptime target with downtime defined and six named exclusions. What is not offered is any warranty on output, since section 6 disclaims accuracy and completeness, and no insurance is mentioned. Verified 2 September 2026.

Nothing establishable on this date states who bears the loss when the product is wrong, and the cause is two different things which the reader should be able to tell apart. The agreements exist and the vendor says so: a Master Services Agreement and a Data Processing Agreement are both listed by name in the trust center's legal section. Neither is published; both sit behind an access request, which is a disclosure choice about how the terms are made available rather than an absence of terms. Separately, the website terms of service linked in the site footer render client-side and returned the hosting platform's shell with no body on this channel, and a search on clause language did not recover the text, so that document is a limit of this reading rather than a vendor decision. The consequence for a buyer is the same either way and is what the grade records: no indemnity, no liability cap, no warranty, no service level commitment, no exclusive remedy and no insurance position can be read before entering a commercial conversation. The exposure worth naming on this product class is that its outputs become fact statements and pleadings a lawyer signs, and its own published failure taxonomy describes seven ways such outputs go wrong, with nothing published about what the vendor stands behind if they do. Trust center read and terms of service attempted 7 September 2026.

Integration is not a feature of this product, it is the product, and the depth is documented accordingly. The systems named are the ones firm work actually lives in: the document management system, email, SharePoint, OneDrive, HighQ, intranets and experience management and metadata systems. What moves is described in both directions. Documents are crawled and indexed in place with no migration and no uploading, so the firm's data remains in its source system; permissions travel the other way, continuously synchronized from those systems into the index; and the firm decides through indexing controls what is crawled at all. What a firm must configure is set out contractually rather than left to implementation: section 1(d) requires continued access to the customer databases named on the order form, requires a named technical contact, and warns that withholding access degrades the service, while the on-premises clause makes the customer responsible for meeting stated hardware and configuration requirements. The platform also exposes governed access to other AI systems through the Model Context Protocol and publishes an open agent handoff specification. Verified 2 September 2026.

One integration, into the system this product class actually runs on, documented on that platform's own marketplace. Fileread is available as a native integration inside Relativity, with a listing on the Relativity app hub, and the integration is described by a customer rather than only by the vendor, a practice support manager at Nutter McClennen & Fish attributing easier adoption to it. That is the deepest integration available in this segment: Relativity is where litigation productions already sit, and the vendor's design position follows from it, that the platform works on the productions a team already has with nothing to migrate. A second distribution surface exists on the Microsoft commercial marketplace, and the trust center lists integrations as a product security control. Cimplifi, an eDiscovery services provider, describes offering the product through its own ecosystem, which is a channel rather than an integration and is recorded as such. What is absent is breadth and documentation. No API or developer documentation was located anywhere in the site navigation or footer. No document management system, case management platform, word processor or e-signature product is named. No integrations page or connector directory exists, and nothing describes what moves in which direction between Fileread and Relativity or how permissions travel across that boundary. Homepage, footer, Relativity app hub listing and trust center checked 7 September 2026.

Three deployment models are defined in the agreement itself, with the consequence of each spelled out, which is the most complete treatment of this axis in the pull. Cloud deployment hosts the services in a cloud hosting location named on the order form. On-premises deployment runs inside the customer's own infrastructure, with DeepJudge granted access to install and maintain, and the customer responsible for stated system requirements. Hybrid deployment is defined precisely as on-premises with certain processing, expressly including model processing, occurring outside the customer's infrastructure, and the clause states plainly that customer data may be processed in the cloud hosting location and that model processing may involve transfer of customer data outside the customer's systems. That is processing separated from storage, stated by the vendor rather than inferred, and it names the one circumstance in which an on-premises customer's data leaves. The security page adds that the customer runs its choice of model in its choice of residency. The limitation worth recording is that no specific regions are enumerated: residency is selectable and designated on the order form rather than published as a list. Verified 2 September 2026.

Neither co-equal limb is stated, and what is published in their place is infrastructure. The cloud providers are named clearly and in more than one place: data is stated to be encrypted and hosted in Azure and processed through private large language models, and the trust center lists both Amazon Web Services and Microsoft Azure under infrastructure and Microsoft Azure again among the subprocessors. Naming the cloud is where a workload runs, not the tenancy model and not the region, and it is credited on the supply chain and stewardship rows rather than here. On tenancy, nothing states whether customers are separated logically or physically, whether a matter is isolated from another matter, or what single-tenant options if any exist; the phrase private large language models points at dedicated inference rather than at tenancy of the data store and is not treated as a tenancy statement. On region, nothing states a country or region of storage or processing, no region selection is offered, and no data residency commitment appears despite GDPR compliance being claimed, which ordinarily brings a transfer position with it. Two gated artifacts in the trust center would answer this directly and were not read, being a Data Flow Diagram and a Network Diagram, and they are the named rebuttal route on this row. Security page, homepage and trust center checked 7 September 2026.

Two attestations are claimed and one carries its scope. SOC 2 Type II is stated as covering security, availability and confidentiality, and ISO 27001 certification is stated as covering the information security management system, which is more than most records in this pull offer. The data processing addendum, version 1.1 of May 2026, records annual penetration testing in Annex 2 and states that controls are audited for SOC 2 Type II. What the top band asks for is absent from every surface a prospect can reach: no auditor is named, no certificate number, issue date, report date or coverage period is published, and there is no trust center or portal. The route to evidence is a customer route rather than a public one. DPA clause 9 sets it out: certifications and audit reports or written questionnaires are made available first, and a customer audit is permitted only where those cannot demonstrate compliance, at the customer's expense, once per calendar year, with no access to systems, hosting sites or infrastructure. That is a real entitlement for a signed customer and not accessible evidence for a buyer still deciding, which is what keeps this at B.

A real trust center with a named artifact inventory, and the reports themselves behind a request whose terms the portal does not state. What is ungated and readable: five compliance positions, being SOC 2 Type 2, ISO/IEC 27001:2022 with the standard version given rather than the bare number, HIPAA, GDPR and CCPA; a control inventory spanning roughly twenty categories from continuous monitoring and product security through access control, endpoint, network and application security to incident response, risk management, business continuity, asset management, change management and training; a named subprocessor list; and a subscribable feed of trust center updates carrying dated subprocessor changes. That breadth of ungated detail is more than most records in this corpus publish at all. What sits behind the access flow: the SOC 2 report, an ISO 27001 report, a penetration test report, a data flow diagram, a network diagram, the policy library and the agreements. The portal offers to start a security review, to view and download sensitive information and to request access, and does not state whether access is granted instantly on an email address, by a self-service non-disclosure click-through, or by vendor approval after a sales conversation. Under the standing rule for gated artifacts the lower tier is taken where the portal does not say which it is, and that is why this is not the top grade, alongside the absence of any named auditor, audit period or scope statement. No request was submitted. Trust center and security page read 7 September 2026.

The original note said that not one provider was named and that no change notification commitment existed. Both are corrected: Annex 3 names Microsoft Ireland Operations Ltd, Google Cloud EMEA Ltd and AWS EMEA SARL, each described as AI provider and cloud infrastructure with location customer selected, and clause 5 commits to at least thirty days notice before any new subprocessor processes customer data, with a right to object on reasonable data protection grounds and termination without liability if the objection cannot be resolved. The grade does not move, and the reason is the shape of what is named. The three entities are hyperscaler platforms rather than model vendors: no discrete model house appears anywhere in the list, no model is named, and no default is stated, so a reader learns which cloud estates the inference runs in but not whose model produced the output. Provider-level naming does not reach the top band and platform-level naming reaches it less. What is genuinely well disclosed is the architecture and the geography: DeepJudge states that its architecture is model-agnostic and that the customer chooses from and may combine any commercial or open-source models, running its choice of model in its choice of residency, and the hybrid deployment clause in the subscription terms discloses that model processing may involve transfer of customer data outside the customer's own infrastructure. Customer-selected location on all three AI providers corroborates the deployment position rather than adding to it.

The fullest provider-level disclosure located in this lane, and the models are still not named. The trust center publishes the subprocessors on its public page rather than behind the document request, and the list is legible as an AI stack rather than as generic infrastructure: OpenAI, Baseten, Pinecone, Langfuse and Microsoft Azure, with Amazon Web Services and Azure also listed as infrastructure. A reader can therefore establish that a named model provider is in the path, that inference serving and vector storage sit alongside it, and that a prompt observability layer is involved, the update feed recording in terms that Langfuse was added as a subprocessor for prompt use and that OpenAI was added as a subprocessor. Change notification exists as a published mechanism rather than a promise, the trust center carrying a dated subprocessor changes feed with a subscribe option, which is a limb most records on this axis fail outright. Two things keep it off the top grade. No model or model family is named: the site says data is processed through private large language models, and under the standing rule naming a provider is not naming the models underneath. And nothing states which parts of the platform route customer content to which provider, so a reader knows OpenAI is in the stack without knowing what it sees. Trust center and security page read 7 September 2026.

The unit and the structure are published in the subscription terms without any figure. Fees sit on an order form and are stated to be in United States dollars unless otherwise agreed. The term structure is set out: subscriptions renew automatically for successive one-year periods unless either party gives three months notice before the end of a term, invoices issue within ten days of the effective date and each subsequent term and are payable within thirty days, and late payment attracts the greater of five percent per annum or the statutory maximum. Professional services are separated cleanly, provided only under a mutually agreed statement of work that becomes part of the agreement, which tells a buyer that implementation is priced apart from subscription. Most usefully, the consumption unit is published: section 4(g) sets a hosting threshold of 500,000 documents per user or 250 gibibytes per user, above which DeepJudge may seek volume reduction or charge at standard rates. What is absent is any figure, any tier, and any pricing page; the only commercial route on the site is Book Demo. Verified 2 September 2026.

No pricing information is published at any level, including the unit of charge, and the site's own navigation establishes that no pricing page exists rather than that one could not be reached. The published inventory of the site is Product, Solutions with three segment pages, Customers, Security and Resources, with a footer adding Workflows, Company, Privacy and Terms of Service. There is no pricing entry in either the header or the footer. Every conversion path on every page ends at the same call to action, an invitation to see the product on your matter, which is a demo request. No tier or edition is named, no unit is identified whether by seat, matter, document, gigabyte or production, no band or range appears, no minimum or term is stated, and no free tier or trial is described. The two documents that would carry the commercial mechanics are unavailable: the Master Services Agreement is listed in the trust center behind an access request, and the website terms of service render client-side and returned no body. Where the only route is an invitation to contact sales, no pricing row is owed and none is written. Site navigation, footer, security page and trust center checked 7 September 2026.

Who buys this is evidenced rather than asserted, and the roster does the work: global elite firms including Freshfields, Greenberg Traurig and Holland and Knight, United States mid-market practices including Cozen O'Connor, ArentFox Schiff and Vorys, a technology-sector specialist in Gunderson Dettmer, Swiss and Austrian firms including Homburger, Lenz and Staehelin, Wenger Vieli, Advestra and Schoenherr, and an in-house legal and compliance function at SBB, so both private practice and in-house use are demonstrated. Geographic reach is stated as teams across North America and Europe serving clients worldwide. Practice areas are not enumerated, and on this product that limb does not bite in the usual way, because the platform searches whatever the firm already holds rather than covering any defined body of law. What is genuinely missing is the boundary: no firm size band is named, nothing states a minimum data estate or user count below which the product is not sold, and no statement describes where it stops. The clearest limits published are the prohibited users at section 4(j) rather than any positive statement of scope. Verified 2 September 2026.

The buyer, the workflow and the material are each described with real specificity, and no boundary is stated. Three buyer segments have their own surfaces: litigation teams, eDiscovery, and corporate legal. Five workflows are published individually, being early case assessment, discovery document review, evidence analysis, deposition preparation, and factual briefs and case narratives, and the lifecycle claim is explicit, from first production to trial. The named customer set matches that spread rather than sitting in one corner of it, running from a Washington litigation firm and a boutique through a large firm's knowledge and practice support function to a litigation administration provider and an eDiscovery services company. Coverage of the material itself is unusually concrete and is the strongest part of this row: text, images, handwritten notes, audio, video and spreadsheets in the same matter, with the vendor claiming the formats that defeat other tools, illustrated by a production that arrived as 4,000 text message screenshots. What is absent is any limit. No jurisdiction is named, and none is excluded, though the customer set and the vocabulary are United States civil litigation throughout. No matter type, case size or practice area is named as in or out of scope. No volume ceiling is published despite scale being the product's central claim. Homepage, solutions and workflow navigation and customer testimonials read 7 September 2026.