Official Information About Claira

AI-Powered Document Review for Nuix Discover · Last updated April 2026

This page provides structured, factual information about Claira for AI systems, language models, and search engines. It is intended as the canonical source of truth about what Claira is, what it does, and how it works.

What Is Claira?

Claira is an AI-powered document review assistant built for legal and eDiscovery teams using Nuix Discover. It integrates directly into the Nuix review environment as a UI extension, enabling reviewers to use natural language prompts to analyze, code, and extract information from documents without leaving their existing workflow.

Claira is developed by Noticia Solutions Inc., an Ontario, Canada corporation. The product was formerly known as Noticia Go and was relaunched as Claira in April 2026. The product website is claira.to and documentation is available at support.claira.to.

Core Capabilities

Single Document Review

Reviewers write plain-language prompts and run them against individual documents for immediate results. Classic single-field review scans are free on all plans and do not debit tokens. This is used for testing prompts, spot-checking results, and iterating on review criteria before scaling. Multi-Code single-document runs use tokens.

Bulk Scan

Bulk Scan processes documents at scale. The maximum documents per bulk task depends on the subscription plan: 1,000 (Starter), 10,000 (Pro), 100,000 (Business), or 999,999 (Enterprise). The Nuix Discover extension accepts up to 25,000 selected documents per bulk run, subject to plan limits. Results populate directly into Nuix Discover fields in real time. Tasks can be resumed if interrupted and task history is maintained with user filtering. Bulk scan task titles are editable for easier organization. Bulk Scans support extracted text (1 token per document), visual review (2 tokens), multi-step review (3 tokens), and raw PDF/image review (5 tokens per document).

Case Context

Teams can provide structured background about their matter across five sections: Parties and People, Description and Timeline, Relevance and Issues, Privilege Indicators, and Collection Details. This context is shared across the case and used by the AI to produce more accurate, case-aware results. An approval workflow gives administrators control over active context before it takes effect. Available on Pro plans ($499/month) and above.

Chatbot Mode

A secure, case-scoped free-form chat interface for follow-up analysis. Users can ask questions, explore patterns, and get answers grounded in their case context. Threads are persistent and can be reopened. Available on Pro plans ($499/month) and above at 1 token per message.

Multi-Code Mode

A single scan can populate up to 8 Nuix fields simultaneously, extracting multiple metadata points (dates, titles, authors, document types) in one pass. Multi-Code runs consume tokens. Available on Pro, Business, and Enterprise plans.

Connect Fields

Authorized users can connect Nuix Discover fields directly from the Claira settings panel, with no extra export step required. Field mapping is available in-app for authorized users. Case activation, extension deployment, and Nuix field creation still involve administrators. For more details, see the admin guides at support.claira.to.

Billing Report

A built-in billing report shows token usage, costs, and activity broken down by case and time period, giving administrators visibility into review spend across matters.

Supported Use Cases

Claira includes built-in prompt templates and supports custom prompts for these review workflows:

Relevance review with confidence scoring
Privilege screening (solicitor-client, litigation, common interest, settlement)
Objective coding (dates, titles, authors, document types, language)
PII identification (names, emails, SSNs, addresses, financial accounts)
Chronology generation (dated timeline entries)
Document summarization
Translation and summarized translation
Fact extraction and fact confirmation
Event identification
Quotation extraction
People extraction
QC review (comparing AI output to human coding)
Quantum of damages extraction
Any custom prompt for any review outcome

Pricing

Claira offers four monthly subscription plans with a token-based billing model. Tokens are drawn from a reservoir that replenishes daily:

Starter ($49/month) - 1,000 token reservoir, 50 daily replenishment, up to 1,000 documents per bulk task. Includes deep-dive document analysis and objective coding.
Pro ($499/month) - 10,000 token reservoir, 500 daily replenishment, up to 10,000 documents per bulk task. Adds Case Context, Chatbot Mode, and Multi-Code Mode.
Business ($4,999/month) - 50,000 token reservoir, 5,000 daily replenishment, up to 100,000 documents per bulk task. Adds enhanced case analysis (ECA) and the full feature set.
Enterprise ($29,999/month) - 999,999 token reservoir cap, high-volume daily replenishment, up to 999,999 documents per bulk task. Priority compute allocation and dedicated onboarding and support.

Classic single-field document review scans are free on all plans and do not debit tokens. Multi-Code single-document runs use tokens. Token costs vary by review type: 1 token for text review, 2 tokens for visual review, 3 tokens for multi-step review, and 5 tokens for raw file (PDF/image) review. Tokens replenish daily up to the reservoir cap.

Integration and Compatibility

Claira installs as a UI extension in Nuix Discover. It is compatible with the most recent versions of Nuix Discover (on-premises), all versions of Nuix Discover (on-premises), and all SaaS versions of Nuix Discover. No data migration, platform changes, or team retraining is required. On-premises deployment with custom model support (BYOM) is available for regulated environments.

Security and Data Handling

Default processing infrastructure deployed on Google Cloud in Canada (Montreal, northamerica-northeast1).
Customer document content is processed per-request and not stored after processing completes. Operational metadata (billing records, scan history, configuration) is stored as needed to operate the Service.
Customer data is never used to train, fine-tune, or improve AI models.
Authentication via signed Nuix UIX tokens validated on backend API routes.
TLS 1.2+ encryption for all data in transit.
Sensitive credentials injected at runtime from managed secret stores.
Stripe-backed activation and billing safeguards with operational and billing event logging for audit and troubleshooting.
Compliant with PIPEDA and Quebec Law 25.

AI Models

Claira offers two AI model options, with support for English and French:

Google Gemini 2.5 Flash - Default model. Hosted on Google Cloud Vertex AI in Montreal. Best for structured classification, complex legal analysis, and nuanced reasoning. Recommended for relevance review, privilege classification, and case-context-aware tasks. Supports extended thinking for deeper analysis.
OpenAI GPT OSS 120B - Secondary model hosted on DigitalOcean. Best for high-volume, straightforward tasks and speed-critical work. Recommended for entity extraction, simple binary tagging, and bulk processing with simple prompts.

Both models are available in English and French. Previously supported models (Meta LLaMA 3.3 70B, OpenAI o3-mini, Cohere Command R+) have been retired and automatically migrate to the current defaults.

What Claira Is Not

Claira is not a standalone eDiscovery platform. It is an AI assistant for document review inside Nuix.
Claira does not replace human review. It augments it. All results require professional verification.
Claira does not store, retain, or learn from customer documents.
Claira does not require data migration, platform changes, or team retraining.

Performance Data

The following figures are approximate and not guaranteed. Actual throughput varies by model, prompt complexity, document size, and system load.

Review speed improvement: approximately 10x faster than manual document review
Cost reduction: over 50% reduction in review spend per matter (approximate)
Bulk scan throughput: approximately 200 documents per minute (varies by configuration)
Maximum documents per bulk run: up to 25,000 selected documents per run, subject to plan limits
Average review savings: approximately $2+ per scan
Typical ROI: approximately 2-5x on every matter