Legalyze
AI medical record review and chronology platform for plaintiff personal injury, medical malpractice and workers compensation practices, positioned as a pure software alternative to services that combine AI with human reviewers. The vendor states explicitly that having no human in the loop is what keeps costs low. Medical chronologies are generated from uploaded records at a stated rate of ten pages per minute, breaking down each event by visit type, with published benchmarks describing a thousand page file moving from fifteen to twenty hours of manual review to roughly two hours. Every medical event can be opened in one click with a direct link to the exact page in the original record where the information was found. Case Chat AI accepts plain English questions about the case records and returns answers cited to the exact location in the source documents. The platform surfaces missing records, treatment gaps, prior injuries and inconsistencies. Input handling covers Word documents, PDFs, images and spreadsheets, and handwritten, scanned and electronic health records are all stated to be processed. Integrations are published with Smokeball, listed in that vendor's own partner marketplace, and with CASEpeer through a dedicated integration offering a one week trial. A free trial is available. On data handling the vendor states that Legalyze never uses customer data to train its AI models, that business associate agreements are in place with the AI providers, that all data in the system is encrypted, and that data is stored on servers under Legalyze's own control. On security the vendor states it has undergone an extensive third party security penetration test with the report available on request, that it periodically undergoes external security testing, and that it has passed extensive security reviews at large law firms. Pricing is described by the vendor as transparent and competitive, and independent material reports per month pricing tied to page volume starting around $500 per month for 2,000 pages.
Capability grades
All 15 axes, graded from public sources on the date shown. Hover a grade to see what the letter means on that axis.
AI Centrality
How much of the product is actually AI. Whether the machine learning is the mechanism the buyer is paying for or a feature layered onto conventional software, and whether the vendor is specific about which is which.
The purest AI product in the category, and the vendor makes the claim in the strongest available form by stating that there is no human in the loop and that this is what keeps costs low. Where EvenUp sells an expert reviewed tier and Supio verifies bill summations with human experts on high value cases, Legalyze positions the absence of human review as the commercial proposition. Everything the product delivers is model output: chronology construction from uploaded records at ten pages per minute, event classification by visit type, page level source resolution, plain English question answering over the case file, and detection of missing records, treatment gaps, prior injuries and inconsistencies. Handwritten, scanned and electronic records are all handled, which is extraction work across three quite different input problems. Remove the models and there is no product. Seventh consecutive A on this axis in this category, and the pattern is now unbroken across all seven plaintiff side vendors.
Citation Accuracy and Hallucination Disclosure
Whether the vendor publishes measured accuracy on citations and assertions, grounds output to primary sources, and says plainly what its system does when it does not know. Legal has a documented public record of fabricated citations reaching filed briefs, so an untested claim of accuracy is not evidence.
Page level grounding applied consistently across both output types, with published throughput and no published accuracy. Every medical event can be opened in one click with a direct link to the exact page in the original record, and Case Chat answers are each cited to the exact location in the source documents, so the citation mechanism covers the conversational surface as well as the generated chronology, which several peers do not. Independent material identifies why that matters in this category, noting source linking to the original page is the detail that counts when opposing counsel challenges a timeline. Held at B because the published numbers are speed rather than correctness: ten pages per minute, and a thousand page file reduced from fifteen to twenty hours to roughly two. Those are benchmarks of throughput. No accuracy figure, no extraction recall, no error rate on handwritten records where extraction is hardest, no evaluation and no test corpus were located. A vendor that publishes its processing rate to the page and nothing about correctness has measured the easy quantity.
Autonomy and Oversight Model
What the system decides on its own, what a lawyer must approve, and whether the vendor documents where the review point sits. A tool that drafts under review and a tool that files without one are different products and different risks.
The oversight model is disclosed by being disclaimed, which is unusual and is credited as disclosure even though what it discloses is an absence. The vendor states plainly that there is no human in the loop, framing it as the reason its costs are lower than competitors that employ reviewers. A buyer therefore knows exactly what they are getting, which is more than a vendor that leaves the question open. What is absent is any account of the verification that replaces it: nothing states whether output is checked by any automated means before delivery, whether a low confidence extraction is marked, what the system does when a page is illegible, or what review the vendor expects the attorney to perform. The product does surface missing records, treatment gaps and inconsistencies, which is checking the records rather than checking the model. Held at C: the model is stated, its bounds are not, and the entire verification burden sits with the firm without the vendor saying so.
Operational and Outcome Evidence
Named, dated evidence that the product works in production at real firms or legal departments. Case studies with figures and identified customers count. Unattributed testimonials and launch announcements do not.
Published benchmarks are specific and no customer is identifiable. Stated: ten pages per minute processing, a thousand page file moving from fifteen to twenty hours of manual review to roughly two, and thousands of records reviewed in hours. Independent material attributes those figures to Legalyze's published benchmarks, which is a vendor source relayed rather than an independent measurement. Two further items carry weight without naming anyone: the vendor states it has passed extensive security reviews at large law firms, which is an assertion about procurement outcomes at unnamed firms, and independent material describes Legalyze as among the most widely adopted tools for personal injury medical chronology work alongside Supio and DigitalOwl. Partner listings with Smokeball and CASEpeer are corroborated on those vendors' own surfaces. Held at C because no law firm is named, no case study exists, the one user quote located is anonymous, and no funding, analyst placement or award was found.
Privilege and Confidentiality Posture
How client confidences are handled: attorney client privilege and work product treatment, segregation of one client matter from another, whether client data trains any model, and what the vendor commits to in writing rather than in marketing.
Infrastructure control is stated with one useful specific and no privilege position exists. Published: all data in the system is encrypted, and data is stored on servers under Legalyze's own control, which is a statement about custody rather than a generic secure cloud reference and rules out the arrangement where a vendor's data sits on infrastructure it does not govern. Business associate agreements are stated to be in place with the AI providers, which addresses the downstream chain for protected health information. Against that: no business associate agreement is described as being offered to customers, which is the instrument a law firm actually needs to execute, and independent guidance in this category names a signed BAA before any protected health information is processed as the first baseline check. No SOC 2 or independent attestation was located, and no treatment of attorney client privilege or work product appears. Held at C on that footing, below the peers at B which carry a customer facing BAA, an attestation, or both.
UPL and Professional Responsibility Posture
Whether the vendor is clear that it supplies a tool rather than legal advice, who its audience is, and how it addresses unauthorized practice of law, competence and supervision duties, and jurisdiction limits. ABA Formal Opinion 512 is the reference point.
Not located. The no human in the loop positioning makes the question sharper than for peers rather than softer: the vendor has removed the intermediate professional review its competitors sell, so the attorney is the only reviewer between generated output and reliance, and nothing published addresses that responsibility. No statement that output is not medical or legal advice, no positioning on the supervising attorney's verification duty over a machine built chronology, and no engagement with any bar guidance. Checked the home page, the medical chronologies page, the competitor comparison pages, the blog material and the partner listings on 29 Aug 2026.
AI Governance and Bias Disclosure
Published governance over model behaviour: who owns it inside the vendor, what is tested before release, and what is disclosed about disparate output across matter types, parties, or populations.
Nothing published about how the models are governed, evaluated or monitored. No AI policy, no model card, no bias or fairness testing, no evaluation methodology, no accuracy monitoring, no drift statement, no named governance body and no ISO 42001. The vendor publishes competitor comparison content ranking platforms on performance and cost efficiency, so it engages the evaluation question commercially and not for its own system. The category risk applies with a specific edge here: the product classifies handwritten and scanned records where documentation quality varies systematically by provider and setting, and no evaluation of performance differences across record types or populations is published. Seventh consecutive D on this axis in this category, and the pattern is now unbroken across all seven records. Checked the home page, the medical chronologies page, the blog material and the partner listings on 29 Aug 2026.
AI Safety and Data Stewardship
Retention, deletion, access control, and what happens to prompts and documents after they are processed. Whether the vendor states its subprocessors and its incident practice, or leaves the buyer to assume.
The most complete stewardship position in this category, and the only one on the index that extends a commitment down the supply chain. Three published statements do the work. First, Legalyze never uses customer data to train its AI models, stated without qualification and not scoped to third party, public or shared models, which distinguishes it from the carve outs recorded on LinkSquares, Eve and Exterro. Second, business associate agreements are in place with the AI providers, which contractually covers protected health information reaching the downstream models rather than leaving that boundary undescribed. Third, all data is encrypted and stored on servers under the vendor's own control. Naming the downstream contractual instrument is what lifts this above its peers: every other record in this category answers what the vendor does with the data and leaves what its model providers do unanswered. Held at B rather than A because it remains policy rather than a customer facing term, no retention period is published, no deletion right is described, and the providers themselves are not named.
AI Liability and Recourse
What the vendor stands behind contractually when its output is wrong. Indemnities, caps, carve outs, insurance, and whether any of it is published or only reachable through a negotiated agreement.
No published position located. Nothing was found on liability for AI output, warranty, service levels or remedy. The exposure is concentrated rather than diffuse on this record because of the no human in the loop model: an omitted provider, a misread handwritten entry or a missed treatment date reaches the attorney with no intermediate check, and a chronology built on an extraction error will read as complete. The product's gap detection surfaces missing records and treatment gaps, which mitigates one failure mode and does not address what happens when the extraction itself is wrong. No published vendor position addresses any of it. Checked the home page, the medical chronologies page, the comparison pages and the site navigation on 29 Aug 2026.
Practice Systems Integration Depth
How deeply the product reaches into the systems legal work already lives in: document management such as iManage and NetDocuments, Word and Outlook, contract lifecycle management, matter management, e-billing, and court filing systems.
Two case management integrations named and both corroborated outside the vendor's own site. Smokeball is published with a dedicated integration generating chronologies directly from documents already in Smokeball, and it is listed in Smokeball's own partner marketplace rather than asserted only by Legalyze, which is independent confirmation the relationship exists. CASEpeer has a dedicated integration page describing upload from CASEpeer into Legalyze, with a one week trial offered to CASEpeer firms. Both are the case management systems plaintiff firms actually run on. Held at B rather than A because the depth is shallow and one directional: both integrations are described as getting documents in, nothing describes finished chronologies returning to the matter file, no API is offered, no authentication or permission detail is published, and only two systems are covered against Tavrn's four plus API access. Checked the Smokeball partner listing, the CASEpeer integration page and the site navigation on 29 Aug 2026.
Deployment Model and Data Residency
Where the software runs and where the data sits. Multi tenant cloud, single tenant, private deployment, on premises, and whether region of residence is a published option or an enterprise conversation.
Custody is stated and geography is not. The vendor states that data is stored securely and encrypted on servers under Legalyze's own control, which answers who holds the infrastructure and says nothing about where it sits. No country, region or data residency commitment is published, no hosting provider is named, and no single tenant or dedicated instance option is described. The statement about server control is credited on the Privilege axis where it belongs and does not substitute for a residency answer here. Compare Supio at B with data centres named in three countries, and ProPlaintiff at C for stating United States storage. Checked the home page, the medical chronologies page, the comparison pages and the site navigation on 29 Aug 2026.
Security Certifications and Trust Center
Independent attestation a buyer can pull without a sales call: SOC 2, ISO 27001, penetration test summaries, a trust center with current reports and named scope rather than a badge image.
A real artifact with a real request route, and no certification of any kind. The vendor states it has undergone an extensive third party security penetration test and that the report is available upon request, that it periodically undergoes external security testing, and that it has passed extensive security reviews at large law firms. Under the three tier test the artifact is a self serve request flow rather than absent, which is why this sits above the D recorded on peers with nothing at all. But the artifact is a penetration test report, which examines technical vulnerabilities at a point in time, and it is not an attestation over a control environment across a period. No SOC 2 of either type, no ISO 27001, no named auditor, no examination scope or date, and no trust centre were located. Independent guidance in this category names a current SOC 2 Type II report as a baseline check, and this vendor does not meet it. Held at C: better evidenced than nothing, and materially short of the peers at B and A which carry named attestations.
Model Supply Chain Disclosure
Which models sit underneath, whose they are, where they run, and whether the vendor commits to telling customers when that changes. A legal buyer inherits every dependency it cannot see.
The existence of the chain is confirmed and contractually covered, and no party in it is named. The statement that business associate agreements are in place with the AI providers does two things at once: it establishes that third party model providers process customer data, which most vendors leave to inference, and it states that the relationship is governed by the instrument HIPAA requires for protected health information. That is more supply chain disclosure than any other record in this category makes, and it is the reason this grades at C rather than D. What is missing is identity: no provider, model family or version is named, no subprocessor list is published, and nothing states how many providers are involved or what each processes. A firm can tell its client that the downstream models are under BAA and cannot tell them who the models belong to. Compare Onspring at B, the only record on this index naming its provider outright.
Commercial Transparency
Whether a buyer can learn what this costs without entering a sales process: published rates, the unit being charged, what sits behind an enterprise tier, and what implementation adds.
The vendor claims transparent pricing, the pages located route to a demo request, and independent material reports a specific published structure, which is a combination worth recording precisely. Vendor comparison pages state that transparent and competitive pricing keeps costs low, and the call to action beside that statement is Request a Demo, with a free trial offered separately. Independent material states that Legalyze publishes per month pricing tied to page volume starting around $500 per month for 2,000 pages, and describes this in contrast to peers using custom quotes, which suggests a published pricing page exists that was not reached in this pass. Graded C on that basis: the pricing model is discoverable at structure level, being per month against page volume, and no figure was located in vendor material. Recorded as a correction candidate in the upward direction, since locating a published pricing page would move this materially. A per page structure is also the most transparent unit in this category for a contingency firm, because case cost scales with record volume rather than being a flat unexplained fee.
Firm and Practice Coverage
Who the product is actually built for. AmLaw, midlaw, small firm and solo, in house departments, government and courts, and which practice areas are supported rather than merely claimed.
Practice scope is stated and input coverage is the most specific in the category, which for a record review product is the coverage question that matters. Practice areas named are personal injury, medical malpractice and workers compensation. Input coverage is enumerated rather than gestured at: Word documents, PDFs, images and spreadsheets accepted, with handwritten, scanned and electronic health records all stated to be processed. Naming handwritten records explicitly is the first instance on this index and it is the hardest extraction case in medical review, where provider notes are often the only record of a symptom. Held at B rather than A because nothing is characterised beyond the list: no statement of whether accuracy or throughput holds evenly across handwritten and electronic records, no jurisdictional scope, and no indication of practice depth in workers compensation as distinct from personal injury, which involves different record profiles and different statutory frameworks.
Legal Signals
What each signal meansA signal records what public sources say on the date shown. It is not a grade and it is not a recommendation. Where a signal reads Not addressed, it means the index did not locate the material in public sources on that date, which is a statement about disclosure rather than about the product.
Client Data in Training
Can material a lawyer puts into this product be used to train a model?
A public policy or trust page states no training on customer content, with no matching term located in the published agreement.
Policy never, and the only unqualified one in this category. The quoted commitment carries no scope limiter: it is not bounded to third party models as on LinkSquares, to shared models as on Eve, or to a named product layer as on Exterro. It is repeated in the vendor's own product material as our AI models do not train on your data. It is reinforced by a second statement that business associate agreements are in place with the AI providers, which extends contractual coverage of protected health information to the downstream model layer rather than leaving that boundary undescribed, and no other record on this index makes that extension. Held at policy never rather than contractual never because both statements appear in product and partner marketing rather than in terms, a data processing agreement or any document a customer signs, and no customer facing agreement was located. Checked the Smokeball partner listing, the medical chronologies page, the comparison pages and the home page on 29 Aug 2026.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
No located public material states how long prompts and outputs are retained.
Not addressed. No retention period is published for uploaded medical records, generated chronologies or Case Chat queries, and nothing indicates whether a firm can delete data or require deletion at matter close. The vendor states that data is encrypted and stored on servers under its own control, which describes custody rather than duration. The gap sits oddly beside an otherwise strong data position: this vendor answers what it does with customer data more clearly than any peer in this category and says nothing about how long it keeps it. Checked the medical chronologies page, the comparison pages, the Smokeball partner listing and the site navigation on 29 Aug 2026.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
No located public material addresses walls or matter level segregation.
Not addressed. No permission model, matter level access restriction or tenant segregation description was located. The Case Chat feature makes the question concrete: it answers plain English questions across a case file, and nothing states whether it is scoped to a single matter or can reach across a firm's uploaded estate, or whether answers respect the asking user's permissions. Integrations exist with Smokeball and CASEpeer, both of which carry their own permission models, and nothing states whether Legalyze inherits them on import. Checked the medical chronologies page, the integration pages, the comparison pages and the site navigation on 29 Aug 2026.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
No located term or policy addresses third party requests for customer data.
Not addressed. No government or law enforcement request clause, no commitment to notify a customer before producing their data, and no transparency report were located. The vendor holds claimant medical records including handwritten provider notes on servers under its own control, and business associate agreements with its AI providers govern permitted uses between those parties rather than constituting a public notice commitment to customers. Checked the medical chronologies page, the comparison pages, the Smokeball partner listing and the site navigation on 29 Aug 2026.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
No located public material identifies the corpus behind the product’s answers.
Not addressed, and inapplicable in the usual sense with a live residue. The platform operates on the firm's own uploaded records rather than a published law corpus, so there is no external legal source to name, license or date. The residue is what the extraction models were trained on, and it is narrowed rather than answered by the vendor's own disclosures: customer data is stated never to be used for training, and business associate agreements exist with third party AI providers, which together indicate the models are supplied rather than built on customer records. Nothing states what those supplied models were trained on or whether any legal or clinical fine tuning corpus exists. Checked the medical chronologies page, the comparison pages and the Smokeball partner listing on 29 Aug 2026.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
No located public material addresses whether authority is checked for subsequent history.
Not addressed, and inapplicable on the facts. The product extracts and organises medical facts and answers questions about the case record. It produces no legal authority and no demand letter drafting was located in vendor material, so there is nothing for a citator to check. Recorded as a scope fact rather than a disclosure failure, consistent with the treatment on Tavrn and DigitalOwl, so a reader comparing this record against a research or drafting product does not mistake an inapplicable row for a gap. Checked the medical chronologies page, the home page and the comparison pages on 29 Aug 2026.
Refusal and Uncertainty Behaviour
What does the product do when the answer is not in the corpus?
No located public material addresses what the product does when it cannot ground an answer.
Not addressed, with one genuine adjacent capability that is credited in the note and does not satisfy the signal. The product surfaces missing records, treatment gaps, prior injuries and inconsistencies, which tells a user where the record set itself is incomplete or self contradictory and is a real and useful form of flagging. It is a statement about the records, not about the model's confidence in what it extracted from them. Nothing published describes what happens when a handwritten page is illegible, whether a low confidence extraction is marked, or whether Case Chat will decline to answer a question the file does not support rather than answering from inference. The no human in the loop model makes this the load bearing question, because there is no reviewer positioned to catch a confidently wrong extraction before it reaches the attorney. Checked the medical chronologies page, the comparison pages and the home page on 29 Aug 2026.
Fabricated Citation Record
Does a public court record exist involving output from this product?
No court order, opinion or disciplinary record naming this product has been located as of the date shown. This is a statement about the public record, not a finding about the product.
None located, with the instrument named. General web searches combining the vendor and product names with court, order, sanction, fabricated citation and medical record terms returned nothing on 29 Aug 2026, and no named docket database or court record tracker was searched. Recorded as a statement about what this search found, not as a clearance. The exposure shape is a misstated or unsupported medical fact rather than a fabricated legal citation, since the product generates no legal authority, and the adverse finding to look for would be a court addressing a chronology that misrepresented the underlying records.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
No located public material engages with bar or ethics guidance.
Not addressed. No named ethics opinion, no ABA Formal Opinion 512, no state bar guidance and no engagement with professional conduct rules was located, despite the vendor publishing comparison and ranking content aimed at attorneys evaluating platforms. That content engages performance and cost and not the professional duties attaching to reliance on the output. Sixth of seven records in this category at this value, with ProPlaintiff the only exception. Checked the home page, the blog material, the comparison pages and the site navigation on 29 Aug 2026.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
Public materials claim time savings without addressing billing or disclosure.
Savings claims only, and unusually the savings are quantified as a rate rather than a percentage. Published: review speeded up by ten times, ten pages per minute, a thousand page file reduced from fifteen to twenty hours to roughly two, and a user report of hundreds of man hours of document review saved. Those describe the firm's own labour. Nothing appears on the claimant's side: no position on whether platform cost is a case expense or firm overhead in contingency work, no disclosure guidance, and no record a firm could produce showing what portion of a chronology was machine generated. Worth noting that per page pricing tied to record volume, which independent material attributes to this vendor, is the structure that would make a case cost most legible to a client, and the vendor does not connect the two. Checked the medical chronologies page, the comparison pages and independent review material on 29 Aug 2026.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
The material exists behind a sales conversation or an executed agreement.
On request, reached through a fourth distinct route in this category. The vendor states that an extensive third party security penetration test has been performed and that the report is available upon request, which is a named artifact with a stated request path. Alongside it, two statements a firm could actually use in a client questionnaire: business associate agreements are in place with the AI providers, and the platform has passed extensive security reviews at large law firms. Held at on request rather than higher because nothing is open and the pack is incomplete: no subprocessor list, no named model provider, no SOC 2 or other attestation, no customer facing BAA described, and no trust centre. Note the category now shows four routes to this value: EvenUp via a Trust Center portal, Tavrn via an offered BAA, DigitalOwl via both, and Legalyze via a pentest report on request.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
Some elements of the record are available, short of a document level export.
Partial record, with the strongest source limb in the category applied across both output types. Every medical event opens in one click to the exact page in the original record, and Case Chat answers are individually cited to their location in the source documents, so a party can show where any assertion came from and an opposing party can check it. Independent material identifies this as the detail that matters when a timeline is challenged. The gap detection capability adds something no peer offers here: surfacing missing records and treatment gaps means a party can also see what the record set does not contain, which speaks to completeness as well as provenance. The familiar two limbs remain absent: nothing indicates that output records which model produced it, and no human verification record is captured, which under the no human in the loop model means there is no verification event to record at all unless the attorney creates one outside the platform.