Irys vs Spellbook: how they compare in 2026
Irys and Spellbook both draft and redline in Microsoft Word for law firms and in house teams, but they start from different work. Irys, which traded as Iqidis until March 2026, organizes work around matters. It researches case law with links to the opinions relied on, runs Cite Check on a brief, analyzes a matter's documents through a knowledge graph, and sends Workbench agents to gather material from email and files. Spellbook starts from the contract. Review proposes tracked edits against a team's playbooks, Benchmarks compares language with market patterns, Associate works across several files, and Autonomous Contract Management routes routine incoming paper for in house teams. Spellbook's iManage connection keeps existing permissions and ethical walls, while Irys runs its own matter roles. Only Irys publishes a price, $299 per seat a month for Irys One, and Spellbook quotes by team size. Their contracts differ on training too. Irys rules it out, while Spellbook may use deidentified customer material to personalize that customer's own features. Litigators who want research and citation checks beside drafting fit Irys. Deal teams working in Word and iManage fit Spellbook.
Which one fits
Choose Irys if
- Your work is litigation as much as contracts. Irys researches case law against retrieved sources, drafts motions and memos, and links each result to the opinions relied on.
- Every brief needs a citation check. Cite Check flags citations it cannot match against its sources and surfaces negative history, and the Thorough mode checks case law as it writes.
- You want the price before the demo. Irys One costs $299 per seat a month or $3,049 a year, Irys Teams $329 per seat for 2 to 25 seats, and a seven day trial needs no card.
- Training must be barred in the contract. The data processing addendum says customer content does not train Irys's or any third party model, and providers run stateless with retention disabled.
- Agents should do the gathering. Workbench agents collect material from connected email and files and return finished documents, leaving any email in drafts until you send it.
Choose Spellbook if
- Contracts are the job and Word is the desk. Review proposes tracked edits against your playbooks inside Word or Google Docs, and you approve them one by one or together.
- Your precedents live in iManage. Spellbook brings selected iManage files into Word and Associate while keeping their existing permissions and ethical walls.
- You negotiate against the market. Compare and Benchmarks set a draft's language against market patterns.
- In house intake is swamping legal. Autonomous Contract Management routes routine incoming contracts by your rules and can approve those that pass every check.
- Research crosses borders. The legal data connection retrieves from open sources such as CanLII, CourtListener, Cornell Law and EUR Lex, with links to the authority.
At a glance
All 15 axes, side by side
The same grid applied to every vendor in the index, graded from public sources. 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.
Irys One researches case law, drafts motions, memos and contracts in a firm's voice, redlines against playbooks, analyzes documents across a matter through a knowledge graph built from the matter's files, and checks citations; Workbench agents gather material from connected email and files and return finished documents. Matter management, the document library and electronic signature support that work; without the models there is little left to sell.
The models are the product. Review, Draft, Ask, Compare and the Associate agent are all generative capabilities delivered through a Word add in that exists to carry them. The company pivoted its whole business to generative AI in 2022, rebranding from Rally to Spellbook. Remove the models and there is no product, only an empty add in.
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.
Research runs against retrieved legal sources, and results link to the opinions and documents relied on. Cite Check checks identified citations against legal source databases, flags those it cannot match, and surfaces available treatment signals such as negative history. The Thorough mode checks case law citations against source as it writes, while the Quick mode is disclosed as not backed by citations. The Responsible AI page says plainly that no mode makes output correct and that Cite Check does not verify every citation. The company has published a run of its research system against a public legal agent benchmark of 1,251 tasks, with an 83.74% criteria pass rate and a 17.75% strict pass rate on all criteria. It also published its departures from the benchmark's method, and released the code and traces. That measures agentic legal work, not the accuracy of Irys One's citations.
Ask can use selected iManage files and return citations to the team's own contracts. The legal data connection retrieves material from named open legal databases such as CanLII, CourtListener, Cornell Law, EUR Lex and regional legal information institutes, with links to the underlying authority; its page claims more than 1,000 databases while its tables list about 108. Their currency follows the source operator's updates. Spellbook's Humans Hallucinate Too study ran Review over 3,019 contracts filed with the SEC in a deliberately conservative mode, without a described human check of its findings. Spellbook publishes no reproducible accuracy evaluation for its own AI, no automatic subsequent treatment check and no abstention threshold. A citation makes an answer inspectable, but checking the quotation or legal proposition against it is still the reviewing lawyer's job.
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 Quick mode is stated not to retrieve or verify authority and never to be the basis for authority cited in a filing; the Standard research mode is to be paired with Cite Check before its citations are relied on; the Thorough mode checks citations as it works. Workbench agents gather material and build documents, but emails are left in drafts and never sent without the user. Drafts arrive as tracked changes, Cite Check returns flags for an attorney to resolve, and an audit trail records what was run, when and in which mode.
Review proposes changes in Word for the lawyer to inspect and approve, either individually or in a group. Playbooks encode the team's instructions, while Associate takes on work across several files with user oversight. Proposed edits can carry the reviewer's identity when accepted, and citation links let the reviewer inspect supporting material. Autonomous Contract Management, generally available since 7 October 2026, applies routing rules that can auto approve contracts, such as those under $50,000 that pass every playbook rule, while flagged items go to a person and every AI edit appears as a tracked change. The terms place responsibility for reliance and verification with the customer. No universal threshold is stated at which Associate must stop or escalate an unsupported conclusion, and no route is published for reversing an approval the system made on its own.
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.
The site quotes Omer Er, Partner at Michelman & Robinson, lists users at Falls Counsel, Cox O'Brien and the Law Offices of Li Yang, and says more than 300 legal teams use the platform. Its outcome figures, 41% fewer hours and turnaround cut from 3.9 to 2.3 days, are attributed to two unnamed legal teams in their first weeks on the product.
Named customer accounts involve Dropbox, Panasonic, KMSC Law, Alturas Capital Partners and Elevare Law. Panasonic's account concerns saving three weeks while building an RFP process, while other testimonials describe drafting capacity and keeping work inside the legal team. Spellbook says more than 5,000 legal teams use it and claims 75 percent time saved on contract work without a method. These examples establish identifiable use rather than only a logo wall. A consistent dated measurement protocol, comparison baseline and workload definition do not accompany the range of savings claims, so the figures are specific customer accounts rather than a reproducible general result.
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.
The Data Processing Addendum and Terms of Service commit that customer content is not used to train Irys's or third party models. It is kept separate for each tenant and user profile, never commingled with other customers, and deleted within 30 days of termination or earlier on request. The terms bind Irys to confidentiality, with prompt notice before any compelled disclosure. Third party models are named, used for stateless inference with retention on the provider side disabled, and bound by terms that bar training. Administrators cannot see a user's personal workspace unless it is shared. The security page presents the platform as designed to preserve confidentiality and support privilege, while the terms leave compliance with privilege obligations to the user. Separation by matter is described on product pages, not in the agreement.
Spellbook's agreement expressly protects privileged information and work product and requires confidentiality for staff and service providers handling customer content. Its iManage integration preserves existing file permissions and ethical walls. Third party models cannot train on customer content and operate under zero retention commitments. Spellbook itself may use deidentified content for customer specific personalization and retain inputs and outputs temporarily for quality assurance, while the terms restrict those uses and access. These permissions are narrower than unrestricted cross customer training, but they are not an across the board ban on learning from client content. Application history also has no single stated in use retention deadline.
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. Where the advice line is not the duty a product raises, the axis is read through the nearest professional duty it does raise: judicial conduct rules and the reviewing duty for products sold only to courts, and the duty to bill for time actually spent for products that draft time entries.
The Terms of Service and a separate Disclaimer state that Irys gives no legal advice, creates no attorney client relationship and is an assistive tool for legal professionals. A Professional Responsibility and Ethics section requires independent judgment, review of material output including authorities, quotations and subsequent treatment, review before anything goes to a court, client or regulator, and firm policies, training and supervision for AI tools. A Responsible AI page explains in plain language why attorney review is not optional, and an AI Rules by State guide sets out disclosure requirements, ethics guidance, sanctions and CLE rules state by state, with sources. No consumer facing surface is published.
Output counts as information rather than formal legal advice under Spellbook's terms, which place verification and appropriate lawyer review with the customer. The product is directed to law firms and legal departments, with review controls inside Word and oversight of Associate's work across files. The agreement also warns that non English use can produce unsatisfactory results. That is a concrete language limitation, but not a tested jurisdictional coverage matrix or a map of which regulated services the software can provide. Human acceptance of a redline remains part of contract work, and an answer from a cited public source is not a substitute for checking its legal application.
AI Governance and Bias Disclosure
Published governance over model behavior: who owns it inside the vendor, what is tested before release, and what is disclosed about disparate output across matter types, parties, or populations.
The Responsible AI page sets out what Irys does and does not claim, how its safeguards are layered and what strong firm practice looks like; policies name Legal & Security at Iqidis, Inc. as their owner, and the security page says the architecture was audited by Guidepost Solutions. No accountable individual, internal governance framework, testing before release, or finding on whether output differs across matter types, jurisdictions or practice areas is published.
Spellbook's terms describe internal AI risk and prohibited use policies, staff education and regular reassessment of ethical and legal requirements. Its security material also cites a CMS Law opinion classifying the product under the EU AI Act and identifying applicable controls. These are concrete governance arrangements beyond a general promise to act responsibly. The product supports verification through citations and user approval of edits, but those controls are separate from evaluating a model before release. The framework names no internal owner, no reproducible release test set and no results on uneven behavior across populations or matter types.
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.
Canceled accounts stay exportable for up to 90 days and are then deleted within 30; the Data Processing Addendum commits to deletion within 30 days of termination or earlier on written request; users can delete data at any time. MFA is enforced for all accounts, with least privilege roles, SSO and SCIM, idle session expiry and an exportable audit trail. A subprocessor policy names every service provider and subprocessor with purpose, location and retention, and commits to 30 days' notice of changes with a right to object. Breaches are notified within 72 hours with set contents and equivalent obligations on subprocessors, and government requests are notified where lawful and challenged if overbroad.
OpenAI and Anthropic process Spellbook requests under zero retention and no training commitments. Spellbook's own terms permit quality assurance copies of inputs and outputs for up to 90 days, restricted to authorized staff, and self service abuse monitoring for up to 30 days unless an investigation or law requires longer. Microsoft identity integration lets organizations enforce their authentication policies, and iManage access preserves source permissions. The agreement covers confidentiality, security incidents and deletion or return of customer material, while the processor register identifies downstream services. Outside those secondary uses, Spellbook gives no single fixed period for how long application history is kept.
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.
The Terms of Service cap each party's liability at twelve months' fees or $100, whichever is greater, but lift the cap for breaches of confidentiality and for gross negligence or willful misconduct. Irys defends customers against US patent, copyright and trademark claims over the service, including output used as delivered, and the Data Processing Addendum adds an indemnity for data subject or regulator claims caused by a security incident in Irys's systems. Irys does not warrant that output is accurate or that verification features catch every issue.
Specified intellectual property claims against the software are covered by an indemnity, alongside service warranties, while the accuracy of generated output and responsibility for customer reliance on it are disclaimed. The general limit is twelve months of fees, with exceptions including confidentiality, security, data incidents and indemnification duties. The agreement includes mutual protection for specified data incident claims and gross misconduct. Beta Features have separate warranty exclusions and a CAD 10,000 cap. These are concrete protections for platform and data risks, but they do not provide an accuracy warranty compensating a client when an accepted clause or legal assertion is wrong.
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.
Irys runs inside Microsoft Word through an add-in on AppSource, and connects on the user's authorization to Google Drive, Gmail, OneDrive, Outlook and Dropbox for import and export, as listed in its subprocessor policy. Product pages also show iManage, NetDocuments, SharePoint and Slack, but its FAQ says document management integrations are in development, and what syncs and what a firm must configure is not documented.
The iManage connection starts from Add from iManage in Ask or an Associate Project and authenticates the user's account once. The same connection then works in Word and Associate. Users select files through the attachment control or project file list; selected content becomes available as chat context or project references without a manual download. Answers cite the imported contracts, and precedents can guide review of incoming paper. Existing iManage permissions and ethical walls govern file access. The connection describes movement from iManage into Spellbook's working context; it does not promise automatic return of revised files to the source repository. Spellbook also runs in Google Docs, connects to Slack, OneDrive, Dropbox, SharePoint and Google Drive, and Autonomous Contract Management takes contracts in from Gmail, Outlook, Slack and Salesforce.
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.
The subprocessor policy says data is processed primarily in the US and EU with regional routing on request, hosted and stored on AWS in the US, and gives each AI provider's processing location, US and EU for the main models. Product pages describe hard tenant isolation with containers for each tenant and each user, and name cloud, VPC and on premises options. What each deployment option involves, and which plans include which, is not described beyond Enterprise being priced to scope.
Customer content is stored and processed in Canada and the United States, Spellbook states, with AWS as its primary cloud provider. Third party model requests are handled under zero retention agreements with OpenAI and Anthropic, while Spellbook itself has separate retention rights for quality and abuse monitoring. The Word add in and Associate connect to shared cloud services, including optional sources such as iManage. The public description identifies storage and processing countries but does not offer a per tier residency selector or a dedicated customer deployment. A source system's location therefore does not confine every later processing step to that same location.
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.
The security page lists SOC 2 Type II, ISO 27001 (the subprocessor policy specifies ISO/IEC 27001:2022), HIPAA, GDPR and CPRA compliance, and an architecture audit by Guidepost Solutions; current reports and certificates are available to customers and prospects under NDA on request. A trust center is linked, and the security page and subprocessor policy publish no auditor, scope or report date.
Spellbook states that it is SOC 2 Type II compliant and provides a trust portal for reports, processor information and compliance resources. Its security program uses encryption, Microsoft identity controls and agreements governing model providers. The public security explanation also describes HIPAA support through business associate agreements where appropriate and a legal opinion on EU AI Act classification. These are different assurances with different scopes. The open product explanation leaves out the current SOC audit period, the auditor and the full service scope. The named SOC standard therefore supports a specific certification claim without treating every resource or regulatory statement as an equivalent independent security audit.
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 subprocessor policy names OpenAI for inference and embeddings, Anthropic's Claude and Google's Gemini on Vertex AI for inference, Mistral AI for OCR, Perplexity AI and Linkup for research lookups and CourtListener for case law and citation lookup, each with processing location and retention, and commits to 30 days' notice of any new subprocessor or material change, with a right to object. It also describes the architecture: most requests are handled by Irys's own knowledge graph, retrieval and orchestration layers, with external models called only for discrete, stateless inference steps. Specific model versions are not listed.
Language models come from OpenAI and Anthropic, under zero retention agreements that keep request and response content in memory for processing rather than persistent storage at those providers. Spellbook describes its models as GPT-5, Claude and other models it does not name, while older pages of its own still cite GPT-4o and GPT-4. AWS is the primary cloud provider, and storage and processing are described in Canada and the United States. The trust portal lists processors and records updates when suppliers change. Spellbook does not map models to features or give exact production versions, and a public processor update history is not an explicit commitment to notify every customer whenever a model version or deployment location changes within the same supplier.
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 pricing page publishes a free plan, Irys Lite and Irys Student at $19.99 a month, Irys One at $299 per seat a month or $3,049 a year, and Irys Teams at $329 per seat a month for 2 to 25 seats, with usage limits by plan and no metered overage on Irys One. Enterprise, for 25 or more seats, is priced to scope, and the enterprise deployment program is not priced.
The number of team members on a license sets the price of the suite, which now lists Review, Draft, Ask, Benchmarks, Playbooks and Associate. Law firm and in house packages both include the Word add in and Associate, training and online support, with dedicated support for teams above ten. The in house offering includes a playbook build service, and academic institutions get free access. A seven day trial is available, with extended trials possible for larger organizations. Subscription amounts remain custom rather than a public per user rate. The terms of 17 September 2026 renew orders for 12 month terms and cap annual increases in subscription fees at 10 percent, and Spellbook may change the terms by posting them, effective after 30 days.
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.
Pages address solo practitioners, small, midsize and large firms, in house teams and insurance defense firms, with use case pages for litigation, corporate and contracts, insurance defense, employment, mass tort and ADA compliance, and law students have their own plan. Citation checking is described against federal and state opinions. The jurisdictions, courts and practice areas it does not support are not stated.
Law firm and in house packages are sold separately, and the principal tools center on transactional drafting, review, negotiation and questions about agreements. Associate extends work across several files, iManage access allows firms to use existing contract precedents, and Autonomous Contract Management, generally available since October 2026, handles routine incoming contracts for in house teams. Industry material covers energy, healthcare, finance, technology, manufacturing and consumer businesses, and academic institutions get free access. The legal data connection adds government and court sources across many regions, though breadth of sources is a different matter from tested competence in every practice. The terms warn about non English results. Spellbook does not list the jurisdictions or specialized legal tasks it leaves unsupported.
The 12 legal signals, side by side
Recorded rather than graded. These are the questions a practitioner has to answer before a tool touches a client matter, and the answers are taken from public material only.
Client Data in Training
Can material a lawyer puts into this product be used to train a model?
The Data Processing Addendum states that customer content is not used to train Irys's own or third party models, and that third party models are used only for stateless inference with vendor caching and retention disabled under commitments not to train. The Terms of Service add that training could occur only if expressly agreed in writing. Usage data is limited to aggregated or anonymized telemetry that excludes prompts, outputs and documents.
The terms prohibit third party models from retaining or training on customer content. Spellbook itself may process deidentified customer material to improve that customer's features and calibrate personalized systems, without using that personalization for other customers. It can also evaluate prompt and system improvements through limited quality monitoring. These own service improvement permissions are distinct from the provider prohibition and do not require a separate opt in in the standard agreement.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
Users can delete their data at any time; Irys's pricing page says every plan includes 90 day retention with automatic deletion; canceled accounts stay exportable for up to 90 days and are then deleted within 30; and a data retention policy describing settings and default periods is available on request. Model provider inference is configured for zero retention, and telemetry excludes prompts and outputs. A setting that keeps nothing inside Irys is not described.
Model providers handle request and response content in memory under zero retention agreements. Spellbook may keep quality assurance inputs and outputs for up to 90 days, and self service abuse monitoring data for up to 30 days, with investigation and legal exceptions. Those are specific secondary uses rather than a stated ordinary lifetime for all application history. Deletion and return provisions are a different mechanism from a configurable zero history setting.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
Access is set in Irys's own role model, maintained by the firm: organization administrators, matter owners and contributors, with matter level team collaboration and shared workspaces per team or practice group. The terms say administrators cannot see a user's personal workspace unless it is shared, and the agreement commits to separation per tenant and per user profile. Permissions are not inherited from a document management system; those integrations are described as in development.
The iManage integration applies existing iManage permissions and ethical walls to file access inside Spellbook. A user authenticates the source account once, then selects permitted files for Ask in Word, Ask in Associate or an Associate Project. Selected material becomes context for answers and review. Spellbook describes this protection specifically for the iManage connection and does not extend the claim to permission inheritance on the other optional storage connections.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
The Data Processing Addendum commits Irys, where legally permitted, to notify the customer promptly of a government request for customer personal data, to challenge unlawful or overbroad demands and to disclose only the minimum required; the Terms of Service commit to prompt notice and assistance before any compelled disclosure of confidential information. No transparency report is published.
Section 4.3(d) of Spellbook's terms, updated 17 September 2026, requires prompt notice of a legal or government demand for customer data and reasonable cooperation with efforts to obtain a protective order or contest disclosure. The confidentiality provisions also address compelled disclosure and protection of privileged material. These contractual steps give the customer an opportunity to respond to a demand. Spellbook does not publish a report counting government requests and their outcomes.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
The subprocessor policy names CourtListener, from the Free Law Project, for case law and citation lookup, and Perplexity AI and Linkup for legal and web research lookups; the pricing page describes citation checking against federal and state opinions. The rights basis for each source, its coverage and how often it is updated are not stated.
The legal data connection names its sources, including CanLII, Cornell Law, CourtListener, EUR Lex, BAILII and national government registers. Spellbook retrieves material directly from each source of record and follows its update cadence. The same workflow can use the customer's own contracts and iManage precedents. The named source list makes jurisdiction coverage inspectable, but does not state an individual license or public domain basis for every external collection or its separate market benchmark data.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
Cite Check checks identified citations against legal source databases, flags those it cannot match, and surfaces available treatment signals such as negative history, for attorney review. Irys states that it is not connected to, or a substitute for, any third party citator a firm licenses, and the terms say a passing check does not establish that all subsequent treatment has been identified.
Citations to public case law, statutes and registers come back from the legal data feature, with currency tied to the operator of each source. Ask, Draft and Review can use that authority within the contract workflow. The feature does not describe a licensed citator or a method for checking subsequent judicial treatment of a case. A current link to the underlying decision is therefore separate from a signal that its legal proposition remains reliable.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
Cite Check flags citations it cannot match against its sources, and the Quick mode is disclosed as not retrieving or verifying authority. No behavior in which the system declines to answer, or says it found no support, is described for research or drafting.
Source citations and selected precedent files provide context for Spellbook's answers and proposed redlines. A lawyer can inspect that material and accept or reject suggested changes. The product does not describe an automatic no answer response, an exposed confidence threshold or a required escalation when available material cannot support a conclusion. A contractual warning about unreliable output assigns verification to the user rather than defining an abstention mechanism.
Fabricated Citation Record
Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?
The AI Hallucination Cases database maintained by Damien Charlotin lists one matter naming Irys or its former name, Iqidis: in a declaration filed on 11 August 2026, counsel for a defendant states that she used the product, believed it was connected to her firm's legal research subscription, and did not check the citations before filing. No court order or finding addressing those citations is on the public record. This is a statement about the public record rather than a finding about the product.
Youssian v Village of Kings Point, 2026 NY Slip Op 26150, Supreme Court of New York, Nassau County, September 9, 2026, addresses counsel's fabricated quotations and misrepresented holdings. The opinion recounts counsel accepting Spellbook AI changes without repeating a citation check. It sanctions counsel $5,000, awards $10,000 in fees and strikes both petitions. The findings concern the lawyer's verification and response to the errors; the product attribution appears in counsel's explanation.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
Irys publishes an AI Rules by State guide that names the governing ethics guidance in each state covered, such as NYC Bar Formal Opinion 2024-5, Texas Ethics Opinion 705 and WSBA Advisory Opinion 2025-05, alongside disclosure rules, sanctions cases and CLE requirements, with sources. It maps state rules rather than the product's features to each opinion.
Spellbook's terms treat output as information rather than formal legal advice and place verification and appropriate lawyer review with the customer. Its product, pricing and security pages name no bar opinion or ethics rule, such as ABA Formal Opinion 512, and no state guidance.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
Irys publishes billing guidance covering hourly, fixed fee, value based and blended arrangements, including how to describe AI assisted work on invoices and how to answer client questions about AI use, with invoice, client disclosure and engagement letter templates provided in enterprise deployments. The platform keeps an exportable audit trail of what was run, when and in which mode. The guidance frames hourly work as freed capacity rather than addressing time billed.
The law firm suite is pitched as a way to improve profitability and the in house suite as a way to retain more work internally. Customer stories describe time savings, including Panasonic's RFP process. The product's per user subscription is separate from a lawyer's charges to clients. A matter level invoice record of AI assisted time or instructions for disclosing those savings on a client's bill are not described in the current pricing offer.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
A current subprocessor policy lists every service provider and model provider with purpose, location and retention, and commits to 30 days' notice of changes. Audit report summaries and certificates are offered under NDA on request. No client facing disclosure material is published on the public site.
Spellbook identifies OpenAI and Anthropic, provides a processor register and explains provider retention and training restrictions. Its security resources address SOC 2 Type II, HIPAA support and an EU AI Act opinion, while the agreement covers privileged information and legal demand notice. Those materials support a firm's client review. They do not form a single client consent pack, and the firm's account must distinguish provider restrictions from Spellbook's own personalization and monitoring permissions.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
An exportable audit trail records what was run, when and in which mode; research citations link to their sources, Cite Check flags are kept for review, and drafts arrive as tracked changes. No export for each document recording the model used, the sources retrieved and the human verification is described.
Word review retains proposed tracked edits, and answers based on contracts or public legal data include source citations. Admin audit logs, added in September 2026, record administrative changes rather than where an AI output came from. Those features preserve portions of the review process and allow a lawyer to check supporting material. An export combining exact model version, retrieved sources and human verification for each generated passage is not described.
The available editing and citation records are therefore partial support for an account of AI use rather than a complete court certification package.
Product updates
The latest logged changes from each vendor, newest first, each linked to its source.
Irys
-
Product / capability
Irys made generally available a Word add in update that links the open document to its Irys matter, so the sidebar answers from the whole matter file and lists the documents, notes and attachments it used.
Source -
Integration / interoperability
Irys One Intake now accepts Slack as a source of work.
Source -
Integration / interoperability
Irys made a two way Clio connector generally available.
Source
In summary
Irys
Irys, a New York legal AI platform, organizes work around matters for solo practitioners, firms of every size, in house teams and insurance defense practices. Irys One researches case law against retrieved sources with links to the opinions, drafts motions, memos and contracts in a firm's voice, and redlines against playbooks. Cite Check flags citations it cannot match and surfaces negative history, and Workbench agents gather material from connected email and files, leaving emails as drafts for the user to send. According to the AI Legal Index, Irys's data processing addendum bars training on customer content, and its subprocessor policy names OpenAI, Anthropic, Google, Mistral, Perplexity, Linkup and CourtListener with 30 days' notice of changes. The main caveat is integration: connections to iManage and NetDocuments are described as in development.
Spellbook
Spellbook puts contract drafting and review where transactional lawyers already work, inside Microsoft Word and Google Docs, for law firms and legal departments. Review proposes tracked edits against the user's instructions and playbooks for the lawyer to approve, and Draft and Ask work from precedents and reference files. Benchmarks compares language with market patterns, Associate carries work across several files and projects, and Autonomous Contract Management, generally available since October 2026, routes routine incoming contracts under rules an in house team sets. According to the AI Legal Index, Spellbook runs on OpenAI and Anthropic models under zero retention agreements, with data stored and processed in Canada and the United States. The main caveat is the license: its terms let it use deidentified customer material to personalize that customer's features and keep inputs and outputs for quality checks for up to 90 days.
Questions buyers ask
How does Irys compare to Spellbook?
Irys organizes legal work around matters, pairing case law research and Cite Check with drafting, playbook redlines and Workbench agents. Spellbook centers on contracts inside Word and Google Docs, with playbook review, market benchmarks, the Associate agent and iManage access that keeps existing permissions. From the AI Legal Index, based on each vendor's own published materials as of November 16, 2026. No vendor pays for placement.
Is Irys the same product as Iqidis?
Yes. Iqidis rebranded as Irys in March 2026, and its main product now sells as Irys One. Spellbook went through a rename of its own, from Rally in 2022, when the company moved to generative AI. From the AI Legal Index, based on each vendor's own published materials as of November 16, 2026. No vendor pays for placement.
What do Irys and Spellbook cost?
Irys publishes its prices: a free plan, Irys Lite at $19.99 a month, Irys One at $299 per seat a month or $3,049 a year, Irys Teams at $329 per seat a month, and Enterprise priced to scope. Spellbook prices its suite by the number of team members on the license and quotes the figure, with a seven day trial and free access for academic institutions. From the AI Legal Index, based on each vendor's own published materials as of November 16, 2026. No vendor pays for placement.
Do Irys and Spellbook train AI on client documents?
Irys's data processing addendum says customer content does not train its own or third party models. Spellbook's terms bar its model providers, OpenAI and Anthropic, from training on customer content, but let Spellbook use deidentified customer material to personalize that customer's own features. From the AI Legal Index, based on each vendor's own published materials as of November 16, 2026. No vendor pays for placement.
How do Irys and Spellbook handle citations?
Irys links research to the opinions relied on, and Cite Check flags citations it cannot match and shows negative history, though Irys says it does not replace a firm's licensed citator. Spellbook links answers to open legal sources and the customer's own contracts, and names no citator. Both sets of terms leave checking citations to the lawyer. From the AI Legal Index, based on each vendor's own published materials as of November 16, 2026. No vendor pays for placement.
Neither vendor paid for inclusion, placement or a grade, and neither reviewed this page before it published. Everything on it comes from public material on the dates shown. How the index grades.
Neither vendor names the exact model version behind each feature. Irys lists document management connections to iManage and NetDocuments as in development, and its SOC 2 and ISO reports come under NDA. Spellbook publishes no price figure, keeps inputs and outputs for quality checks for up to 90 days, and describes permission inheritance only for its iManage connection.