S
Saga
Saga is a legal AI platform from Saga International B.V. in Amsterdam, with Norwegian roots, used by more than 500 legal teams and law firms in over 25 countries. Lawyers work in Projects that gather case documents, case law and regulations. Research draws on case law, legislation and regulation from 44 jurisdictions, refreshed every 24 hours, including content from publishers such as Lovdata, Rettsdata and NL Fiscaal.
Every answer links to its source, and Saga says so when no reliable source exists. Grid Review analyzes large document sets in a table, assistants can be built to challenge an argument or play a judge, and a workflow builder sets how much agency each workflow gets. A Word add in reviews and edits against playbooks, an Outlook add in drafts email and files it to a Project, and whole documents translate into more than 50 languages with their formatting kept.
The platform is model agnostic and offers current models from OpenAI and Anthropic. Saga Amplify adds onboarding, training and legal engineering. Solo costs €135 a month on a yearly term, and firm plans are quoted.
Saga, head to head
Side by side on the same 15 capability axes and 12 legal signals: where the two part company, and what neither one publishes.
Alternatives to Saga
Vendors in the same categories as Saga whose graded profiles look most alike, with a line on what each product does.
Similar is not the same, so check that each one does the job you need. No vendor pays for placement.
- CCORTOCORTO is an AI workspace for law firms from CORTO Pty Ltd in Sydney, which also runs a US company, CORTO AI Inc., and an office in San Francisco.
- FFlankFlank runs supervised AI agents that take over high volume legal work for enterprise in house teams, delivered by a forward deployed team rather than as a software rollout.
- LLucioLucio is a legal AI workspace built by lawyers at Zenlegal Technology Private Limited in Bengaluru, which contracts outside India through Lucio Inc and has users in 13 jurisdictions.
- OOrdalieOrdalie is a French legal AI platform for lawyers in private practice, in house legal teams and notaries.
- QQuillioQuillio, formerly AI Legal Assistant, is an AI assistant for Australian and New Zealand lawyers from AI Legal Assistant Pty Ltd, which rebranded the product on 16 June 2026.
- PPhilips SpeechLivePhilips SpeechLive is a cloud dictation, speech recognition and transcription platform sold by Speech Processing Solutions, with a legal edition, SpeechLive Legal, for solicitors…
Ordered by how closely the graded profiles match across the 15 capability axes, inside the same categories. See how each axis is graded in the methodology.
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.
Each part of Saga runs on a model. Chat answers questions about documents and the law, drafts and analyzes contracts. Assistants carry custom instructions, from templates or from scratch, and can challenge an argument or simulate a scenario with a judge or a devil's advocate. Grid Review reads hundreds of documents into a table that sorts by date, number or true and false, for due diligence and mass analysis. The workflow builder chains these blocks, either by drag and drop or from a plain description, and lets the firm decide how much agency a workflow has.
Research, translation of whole documents and drafting from the case file sit beside them. Projects and sharing organize the work, but no matter management or billing system stands apart from the AI.
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 answers link every point to its source. The legal research collection holds about 15 million documents, case law, legislation and regulation for 44 jurisdictions, refreshed every 24 hours, gathered by Saga itself and from partners and publishers such as Lovdata, Rettsdata and NL Fiscaal. Semantic retrieval returns results with the court, date and case number, a cited source is marked as verified, and results can be added to a Project in one step.
Where no reliable source exists, the answer says so instead of offering an invented authority. No measured accuracy rate, test set or failure analysis is published.
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.
Firms decide how much agency each workflow gets in the workflow builder, from fixed reusable steps to agentic runs. In Word, Saga works through redlines and comments the lawyer can accept or reject. Ready made workflows ask the user to confirm the parties before generating letters, notices and reports from the same source data. Summaries, timelines and issue maps carry source references so the lawyer knows where to read in full.
Saga positions judgment, confidentiality and responsibility as staying with the lawyer, and its Amplify legal engineers train firms and build workflows with them before and after rollout. No published threshold makes a workflow stop and hand back, and recovery after a wrong output is not described.
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.
More than 500 legal teams and law firms in over 25 countries use Saga, and the company reported passing €5.5M in annual recurring revenue in September 2026. Named firms are quoted by partners and managers: Nysingh, La Gro, Sherrards, Arcagna, Brækhus, DVDW, Advokatfirma Lippestad, Stek, Poelmann van den Broek and Brinkhof. DVDW describes launching Saga for legal research and document analysis, Nysingh describes its use as a law and notary firm, and Brækhus calls it central to its own transformation.
Saga also announced a partnership with the Brazilian firm Madrona Advogados and reported ranking third among European legal AI assistants used by legal teams in June 2026. No customer publishes a figure for time saved or quality, and no case study with measured outcomes is published.
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.
Customer data does not train models, and the customer remains its owner. Data is kept confidential and segregated from other customers, access is managed between user groups inside a firm, and private links and encryption protect it in transit and at rest. Abuse monitoring and content filters from the model labs are switched off, so prompts are not held by those labs for review. Sole practitioners are promised that confidentiality stays with them while the product does the groundwork.
These commitments sit on Saga's security page and trust center. The service agreement and data processing agreement that govern customer data are not public, and privilege and work product are not addressed by name.
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.
Buyers include law and tax firms, legal departments, sole practitioners and public bodies, among them courts and police organizations. The material keeps judgment, confidentiality and responsibility with the lawyer and presents the product as groundwork rather than advice, with drafts started from the file for the lawyer to finish. Sole practitioners are told the product is safe to rely on because it says when no source exists.
No terms of service are published, so no written disclaimer on legal advice, user eligibility or jurisdiction limits can be read. Professional rules for the bars Saga serves are not discussed, and use by courts is not tied to judicial conduct rules.
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.
Alignment with the EU AI Act is claimed, and an EU AI Act compliance document sits in the trust center behind an access request. The Amplify program offers client firms governance aware adoption, workshops on prompts, assistants and workflows, and quarterly reviews. The trust center groups more than 90 controls under infrastructure, organizational, product and data protection headings, with an internal audit program.
No AI policy, release testing regime, accountable owner or analysis of uneven output across jurisdictions or languages is published.
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.
Subprocessors are listed with their role and region. Microsoft Azure hosts in the Netherlands, Switzerland or the United States, and Weaviate holds the database in the same regions. Amazon Web Services runs models in Germany or the United States, Google Cloud runs models in the Netherlands or the United States, and DeepL translates. Data is encrypted in transit and at rest over private links, and access is managed by user group.
External and internal security tests run regularly, with reports dated December 2025 and February 2026, and continuity, disaster recovery, capacity and configuration management sit among the published controls. Account data is kept while the account is active and for 30 to 90 days after deletion. No retention period for documents, prompts or outputs and no breach notice timeline are published.
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 terms of service, service agreement or limitation of liability is published. The privacy policy refers to a service agreement and a data processing agreement between Saga and each customer, and the data processing agreement is listed in the trust center behind an access request. Caps, indemnities, warranties and recourse for wrong output cannot be read before signing.
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.
A Word add in reviews and edits documents from a prompt, an email or a playbook, fills templates and marks changes as redlines and comments. An Outlook add in, launched in June 2026, drafts and summarizes email, attaches Saga documents and imports emails with attachments into a Project. Saga announced an iManage connection in August 2026, links to the Dutch document systems Epona365 and DMSforlegal in June 2026, and a September 2026 partnership with Dillien for deal data.
Research results can be added straight into a Chat or a Project, and translated documents keep their comments and redlines. Single sign on is set up during onboarding. What each document system connection syncs, and in which direction, is not described.
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.
Cloud hosting runs on Microsoft Azure, with regions in the Netherlands, Switzerland or the United States and the vector database in the same three regions. Model processing is listed separately from storage, on Amazon Web Services in Germany or the United States and on Google Cloud in the Netherlands or the United States. Each customer's data is segregated from the others and reached over private links, and single sign on is configured per firm at onboarding. Translation runs through DeepL. The region a given plan uses, and any on premises option, are not stated.
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.
ISO/IEC 27001:2022 is held. The certificate, the audit report and the statement of applicability sit in a trust center beside dated security test reports, a security training attestation and more than 90 controls, and documents are released through an access request. SOC 2 is pending, and GDPR compliance is stated, with an accessibility statement and the data processing agreement among the agreements on file. The certification body and the certificate dates are not shown on the public view.
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 platform is model agnostic. Saga picks a model for each query, or the user chooses one, and new models are announced as they arrive, such as GPT-6 in September 2026, Claude Opus 5.5 that month and GPT-5 in 2025. Models run through Amazon Web Services in Germany or the United States and Google Cloud in the Netherlands or the United States, both listed as subprocessors. Saga describes the product as a glass box rather than a black box. No commitment to give notice before a model or provider changes is published.
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.
Solo, for sole practitioners and small teams, is published at €135 a month on a 12 month term or €150 a month on a 3 month term, with a 14 day free trial that needs a card. It includes drafting, review, research and workflows across 15 million documents in 40 jurisdictions, with no setup fees and no per document charges, and prices can be shown in euros, dollars, kroner or krona. Saga Enterprise for mid sized and large firms, legal departments and public bodies is quoted after a demo.
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.
Customers range from sole practitioners and small teams to mid sized and large law and tax firms, legal departments, and public bodies such as municipalities, courts, tax authorities and police organizations. Research covers 44 jurisdictions, from the EU, the Netherlands, Norway and the UK to the United States, Brazil and Mexico, and the interface runs in English, Dutch, Norwegian, Italian, German and Turkish. Translation handles more than 50 languages for cross border files.
Ready made workflows cover matter types such as letters, notices, reports and wills, and tax firms use it for tax research. Practice areas and jurisdictions outside the product's range are not stated.
6 public documents
The public pages on file for Saga, with the recorded signals each one supports and the date it was last read. Open any of them and check the reading against the record.
-
Refusal and Uncertainty Behavior, Bar Guidance Alignment, Billing and Fee Posture
Read Oct 13, 2026
-
sagalegal.io/platform2 signals
Primary Law Corpus Provenance, Court Disclosure Support
Read Oct 13, 2026
-
sagalegal.io/security2 signals
Client Data in Training, Ethical Walls and Matter Segregation
Read Oct 13, 2026
-
trust.sagalegal.io1 signal
Outside Counsel Guideline Readiness
Read Oct 13, 2026
-
Fabricated Citation Record
Read Oct 13, 2026
-
sagalegal.io/privacy-policy1 signal
Third Party Request and Subpoena Notice
Read Oct 13, 2026
€135 a month
- One lawyer or a small team can buy Saga Solo.
- Solo costs €135 a month for a year, or €150 a month for three months.
- You can try it free for 14 days, but you need a card.
- Bigger firms and public bodies get a price after a demo.
Solo is one plan for sole practitioners and small teams, with drafting, review, research and workflows over 15 million documents in 40 jurisdictions. It costs €135 a month on a 12 month term or €150 a month on a 3 month term, after a 14 day free trial that needs a card. Prices can also be shown in dollars, Norwegian kroner and Swedish krona. Firms, legal departments and public bodies buy Saga Enterprise on a quote.
Implementation: None on Solo. Enterprise onboarding, training and workshops through Saga Amplify are not priced publicly.
Note: There are no setup fees and no per document charges on Solo. The trial asks the user to install the Word add in and run one workflow on a live matter within 14 days. Enterprise terms, including onboarding and the Amplify training program, are not priced publicly.
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.
Customers remain the owners of their data, and it is not used for training, a commitment made on Saga's security page and repeated as a control in its trust center. Abuse monitoring and content filters from the model labs are switched off. The service agreement and data processing agreement are not public, so no matching contract term can be read.
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.
Core account data is kept while an account is active and for 30 to 90 days after it is deactivated or deleted. With abuse monitoring from the model labs switched off, prompts are not held by those labs for review. How long Saga itself keeps documents, prompts and outputs in Projects and Chats is not stated.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
Segregation is asserted in public materials with no published detail on how it is enforced.
Each customer's data is segregated from other customers, and access inside a firm is managed between user groups. Projects hold the documents, case law and regulations for a matter and can be shared with colleagues. How permissions on a Project are set, and whether they follow a document system such as iManage, is not described.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
Published terms or policy address disclosure to authorities or in response to legal process, and no commitment or reservation regarding customer notice is located anywhere. The vendor has told the customer that data can leave and has said nothing about whether the customer hears of it.
The privacy policy, last updated on 13 August 2026, lists competent public authorities among the recipients of personal data where the law requires it. It makes no commitment to tell the customer about a request or to resist one, and no transparency report is published. Customer data under the service agreement is governed by a data processing agreement that is not public.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
Sources are identified without stating the license or rights basis.
The collection holds about 15 million case law, legislation and regulation documents for 44 jurisdictions, refreshed every 24 hours. Part of it is gathered directly and the rest comes from partners, with the publishers Lovdata, Rettsdata and NL Fiscaal named among them. Coverage runs from the Court of Justice of the EU and national courts in the Netherlands, Norway, the UK and Germany to the United States, Brazil and Mexico. Which sources serve each jurisdiction, and the rights basis for each, are not listed.
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.
Cited sources carry a verified label and show the court, date and case number, and legislation is refreshed every 24 hours. A verified label confirms that the cited document exists. No flag for a decision that has been overturned or a provision that has been amended is described.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
The vendor describes refusal or abstention behavior in public materials.
Answers come from real, cited sources, and where none exists the product says there is no reliable source rather than inventing an authority. The research view marks such a citation as no source found beside the verified ones. No published evaluation shows how often this happens.
Fabricated Citation Record
Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?
No court order, opinion or disciplinary record addressing fabricated or hallucinated legal citations produced by this product has been located as of the date shown. This is a statement about the public record on that one subject, not a finding about the product, and this signal is not a litigation history.
The AI Hallucination Cases database maintained by Damien Charlotin, which records court decisions worldwide that address hallucinated AI content and the tool involved where known, has no entry naming Saga or Saga International. Saga links every research answer to its source and marks a missing source as such.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
Public materials refer to professional responsibility in general terms without naming guidance.
Judgment, confidentiality and responsibility stay with the lawyer in Saga's material, and its Amplify training is described as helping firms adopt AI responsibly, with professional development points available. No ethics opinion or bar guidance from the Netherlands, Norway or any other market it serves is discussed, and no mapping of the product to professional rules is published.
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, and the product sits inside a fee relationship between a lawyer and a client where those savings would change the bill.
Law and tax firms are the main buyers, and the product is pitched on understanding a case in minutes and researching in seconds, while customer partners describe working more efficiently. Saga's own commentary argues that usage based pricing suits large firms rather than smaller ones. No guidance on billing clients for AI assisted work or disclosing it is published, and no record of AI work by matter is described.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
A current subprocessor or model provider list is published.
Subprocessors are listed with role and region on the security page and in the trust center: Microsoft Azure, Weaviate, Amazon Web Services and Google Cloud for model inference, and DeepL for translation. The data processing agreement, the ISO report and an EU AI Act compliance note are released through an access request. No client facing disclosure pack is published.
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.
Every research answer links to its source, Projects keep the documents and authorities behind a piece of work, and the user can see or choose which model answered. Word edits arrive as redlines and comments, and summaries and timelines carry references back to the source documents. No export records the model, the sources and the person who verified a filing, and no court disclosure template is published.