The bridge between AI and your actual documents
Other tools generate text you paste back. Sphere edits your actual DOCX. Tracked changes your counterparty accepts or rejects in Microsoft Word. Research across six legal databases. Run playbooks that encode your organization’s expertise. Build a searchable knowledge base across 15+ document formats.
AI tools that destroy your documents
Legal professionals spend enormous time on repetitive document tasks. Generic AI tools make it worse.
Broken formatting
ChatGPT, Claude, and Copilot destroy numbering, styles, tables, and tracked changes. You spend more time fixing than editing.
No data sovereignty
US-hosted tools fail Swiss and EU client requirements. Your confidential documents leave your jurisdiction.
No audit trail
No record of what the AI did and why. Regulators want evidence. You can’t provide it.
Generic AI
No legal domain expertise. No firm-specific knowledge. No access to legal databases. No playbooks.
Agentic editing with full tracked changes
Tell the AI what you want in natural language. “Replace Acme Corp with TechStart throughout and cap liability at CHF 500,000.” Every edit appears as a Word-native tracked change: strikethrough for deletions, underlined for insertions. Open the DOCX in Microsoft Word and you see exactly what changed.
You can also edit both the source document and the output directly in Sphere. No round-trip to Word required. Manual edits and AI edits coexist in the same tracked changes timeline.
Atomic OOXML operations preserve every formatting detail: paragraph numbering, table structures, header/footer content, styles. The AI uses stable paragraph IDs. It can insert after paragraph 7 without breaking paragraph 12.
Full version history. Every AI agent run creates a new document version with a complete audit trail: prompt, reasoning chain, tool calls, and resulting changes. Roll back to any previous version at any time.
- Natural language editing commands
- Batch staging with review before commit
- Table row insertion matching cell structure
- Image insert, replace, and remove
- Header/footer scoped editing
- TOC maintenance with outline normalization
- In-app editing of source and output documents
- Full version history with rollback
- Template variables: {PARTY_A}, {DATE}
- Download DOCX with tracked changes
This Master Services Agreement (“Agreement”) is entered into by and between Acme Corp (“Provider”) and...
WHEREAS, Provider maintains offices at Acme Corp headquarters in Zurich, Switzerland...
3. LIABILITY.
The liability shall be unlimited.
4. INDEMNIFICATION.
Acme Corp shall indemnify and hold harmless the Client against any claims arising from a breach of this Agreement.
Your expertise, running on autopilot
Playbooks are repeatable AI workflows that can do anything the agent can do. Edit documents with tracked changes. Add review comments. Search legal databases. Generate clause variants. Extract structured data. Fetch web content. Run deep research. If the AI can do it in a conversation, a playbook can do it on autopilot.
Two engines. Extraction playbooks return structured data: parties, dates, financials, risk flags, document classification. Agent playbooks run the full tool suite: read the document, search six legal databases, find clause alternatives, insert tracked changes, add comments with citations. Same capabilities as an interactive session, zero human interaction required.
Auto-Playbooks run the moment a document lands. Configure per folder with inheritance, or library-wide. Stack multiple playbooks: classify on upload, then risk-flag, then extract financials. Results stream in as each completes. Trigger manually, on upload, or programmatically via the REST API.
Playbook prompts are protected. Team members use your organization’s expertise without seeing the underlying analysis logic. Full version history with rollback. Create new playbooks via natural language in the chat, or let the agent create them for you mid-conversation.
- 40+ agent tools available to playbooks
- Extraction engine: structured data output
- Agent engine: full tool suite, headless execution
- Auto-Playbooks: folder-level, stackable, inherited
- Trigger: manual, on upload, REST API
- GDPR DPA, NDA, SLA, IP, Financial, Risk, Classifier
- Protected prompts with version history + rollback
- Create or compose playbooks via chat
Swiss, EU, and ECHR law at your fingertips
Ask anything in natural language. “What are the notice requirements for terminating a commercial lease in Zurich?” “Does EU AI Act apply to contract generation tools?” Sphere queries six legal databases in parallel and returns relevant statutes, case law, and commentary with full citations.
Exact citations work too. Search “SR 220” for the Swiss Code of Obligations or “C-311/18” for the Schrems II ruling. Deep Research mode goes further: it formulates sub-queries, cross-references across jurisdictions, and synthesizes structured research memos.
Because Sphere is fully agentic, research isn’t a separate step. The agent searches law while editing your contract, cites precedent while drafting a new clause, and references regulation while running a compliance playbook. Research flows into tracked changes, comments, and clause variants in the same operation. Encode it into Auto-Playbooks and every incoming document gets research-backed analysis without anyone asking.
Full statute text retrieval. Natural language or SR number lookup.
BGE citation retrieval and topical search across federal decisions.
Swiss legal commentary, data protection blogs, criminal law analysis.
Full regulation and directive text. Natural language or CELEX number lookup.
Case number retrieval and topical search across Court of Justice rulings.
Application number lookup, article-based search across ECHR judgments.
For complex questions, Sphere’s research agent formulates multiple queries, retrieves and cross-references results across all six databases, and produces a synthesized report with full citations. Takes 1-3 minutes for thorough analysis.
Reusable clauses with jurisdiction tagging
Save clauses you use repeatedly: indemnity provisions, liability caps, confidentiality terms, non-compete restrictions. Tag each clause by type and jurisdiction. Hybrid search combines semantic understanding with keyword matching and surfaces frequently-used clauses first.
Building the bank is agentic too. Ask the agent to extract every indemnity clause from your document library, or save the liability cap from the contract you’re reviewing. The agent reads the source, tags the clause by type and jurisdiction, checks for duplicates, and adds it to the bank. No manual copy-paste.
During negotiation, ask the AI to generate 2-4 clause variants at different favor levels, from strongly protective to balanced to concessive. Insert any variant into your current document as a tracked change with a single command.
- 9 types: Indemnity, Liability, Confidentiality...
- Jurisdiction tags: 🇨🇭 Swiss, 🇪🇺 EU, 🇬🇧 UK, 🇺🇸 US
- Hybrid search: semantic + keyword + usage boost
- Agent extracts clauses from docs and library
- Duplicate detection before saving
- AI clause variant generation (2-4 alternatives)
- Usage tracking surfaces popular clauses
Secure cross-org sharing with AI-powered due diligence
M&A transactions, regulatory filings, cross-border deals. Every scenario demands secure document exchange with external parties. Sphere Data Rooms combine granular access controls with automatic AI extraction, so uploaded documents are classified, analyzed, and risk-flagged before anyone opens them.
Invite individual external participants by email (no DocIQ account required) or add entire Sphere organizations for cross-org collaboration. Every access, download, and view is logged in an immutable audit trail.
- 4 roles: Viewer, Collaborator, Editor, Admin
- Invite individuals or entire Sphere orgs
- Per-user folder access restrictions
- Permission toggles: download, print, upload
- Auto-extraction playbooks on upload
- Immutable audit trail for every action
- M&A due diligence, regulatory filings
- Cross-border transaction support
Your documents become a searchable knowledge base
Every uploaded document is automatically indexed for RAG-powered search. “Which contracts have uncapped liability?” “Find all NDAs expiring before July 2026.” “Which supplier agreements lack force majeure clauses?” Hybrid search combines semantic understanding with keyword matching. Answers cite the exact document and paragraph.
Hybrid Search
Vector (semantic) + BM25 (keyword) with Reciprocal Rank Fusion. Results ranked by relevance with direct citations.
Auto-Processing Pipeline
Upload → text extraction → chunking → embedding → indexing. Documents are searchable within seconds of upload.
15+ Formats
DOCX, PDF, DOC, RTF, XLSX, CSV, PPTX, ODP, Markdown, and more. Full text extraction across all supported formats.
AI Classification
The Document Classifier playbook automatically categorizes uploaded documents by type, saving manual tagging effort.
Tags & Folders
Color-coded tags, hierarchical folders, visibility controls (private or org-wide). Bulk operations for multi-select.
Confidence Indicators
Visual confidence bars with exact percentages. Analysis detail panels show all findings, flags, and summaries.
Process entire libraries and data rooms automatically
Upload 500 documents to a data room. Auto-Playbooks classify every file, extract key terms, flag risk clauses, and surface compliance gaps. No one clicks a button. Results stream in as each document completes.
Run extraction across your entire library: “Extract all limitation of liability caps across our supplier agreements” or “Flag every contract missing a GDPR data processing clause.” Results appear as structured findings with direct paragraph citations.
Works across all indexed formats: DOCX, PDF, XLSX, PPTX, TXT, Markdown, and more. No manual tagging. No per-document clicks. Just results.
- Auto-extract key terms from every upload
- Bulk risk assessment across 100s of documents
- Compliance gap analysis at library scale
- Cross-document comparison and deduplication
- Structured findings with paragraph citations
- Works on DOCX, PDF, XLSX, PPTX, TXT, MD
Everything Word can do. Via AI. With full control.
Insert a new clause after Section 3. Replace a liability cap in paragraph 7 without touching paragraph 12. Delete a redundant provision. Add a comment flagging a risk. Every action the AI takes appears as a standard Word tracked change or redline comment that your counterparty can accept, reject, or reply to in Microsoft Word.
Changes are batched and staged for your review before anything is committed. See every insertion, deletion, and replacement in one diff. Accept the batch, roll back individual operations, or undo everything. You stay in control of what ships.
Under the hood, Sphere operates at the OOXML level: paragraph numbering, table structures, headers, footers, styles, and images all survive intact. No format loss. No copy-paste artifacts.
- Insert, replace, delete at paragraph level
- Tracked changes and redline comments
- Batch staging: review all changes before commit
- Roll back individual edits or the entire batch
- Tables, images, headers, footers preserved
- Full version history with every agent run
This Employment Contract is entered into between the Employer and the Employee as defined in Schedule A.
The Employee shall commence employment on the agreed start date and fulfill duties as outlined in the job description.
The employee shall not engage in any competing activity for a period of 24 months after termination of employment.
Compensation shall be paid monthly in arrears to the Employee's designated bank account within Switzerland.
Either party may terminate this agreement with 3 months' written notice, subject to the provisions of Art. 335c OR.
Upon termination, the Employee shall return all company property, documents, and access credentials within 5 business days.
All information received during employment shall be kept strictly confidential. [Duplicated from Section 4.2]
This contract shall be governed by and construed in accordance with the laws of Switzerland, Canton of Zurich.
Generate clause variants at any favor level
Negotiating a liability cap? Ask DocIQ Sphere to generate 4 variants, from strongly protective to concessive. Each variant includes jurisdiction-specific language, proper legal citations, and a favor-level indicator so your team can quickly compare positions.
Pick a variant and Sphere inserts it into your document as a tracked change. Save the ones you like to the clause bank for next time. The output is a standard .docx with native redlines, ready to send to your counterparty.
- 4 favor levels: Protective → Concessive
- Jurisdiction-aware language (CH, EU, US)
- Side-by-side comparison with favor meters
- One-click insert as tracked change
- Save preferred variants to Clause Bank
- Generate custom variants via prompt
Draft new documents from a single prompt
Describe what you need in plain language. A Swiss employment contract with IP assignment clauses. An NDA for a cross-border partnership. A board resolution. A regulatory filing. DocIQ Sphere generates a properly structured .docx with professional formatting, automatic numbering, and jurisdiction-appropriate language.
Not a template fill. Sphere draws on frontier model knowledge, six legal databases, your clause bank, and your document library. A Swiss employment contract gets Art. 340 OR non-compete provisions. An EU data processing agreement gets GDPR Art. 28 requirements. The output reflects the legal framework you specify.
Every generated document is a real .docx you can open in Word immediately. Edit it with Sphere’s tracked changes, run a playbook review against it, or hand it to your counterparty as-is.
- Natural language → structured DOCX
- Any document type: contracts, memos, filings, policies
- Jurisdiction-aware: draws on 6 legal databases
- Clause bank and library integration
- Professional formatting + automatic numbering
- Immediate editing with tracked changes after creation
Template variables for instant customization
Define variables like {PARTY_A}, {DATE}, {AMOUNT}, {JURISDICTION} in your templates. The AI detects all variable placeholders, presents them in a clean panel, and applies all replacements as tracked changes with a single click.
Every replacement is a Word-native tracked change. Your counterparty sees exactly what was filled in and can accept or reject individual changes.
- Auto-detect {VARIABLE} placeholders
- Fill all variables in one panel
- Apply as tracked changes for transparency
- Save variable sets for reuse
- Custom variable formats supported
- Bulk apply across multiple documents
This Consulting Framework Agreement (“Agreement”) is entered into by and between {PARTY_A} (“Provider”) and {PARTY_B} (“Client”), effective as of {DATE}.
The total compensation shall not exceed {AMOUNT} for the initial term of {TERM_MONTHS} months.
This Agreement is governed by the laws of {JURISDICTION}, Switzerland.
Multi-step legal research with full synthesis
For complex questions, Sphere’s research agent doesn’t just search. It reasons. It formulates multiple sub-queries, searches across all six databases, cross-references results, and synthesizes a structured research memo with full citations.
Watch the pipeline execute in real-time: Fedlex for statutory law, BGE for case law, EUR-Lex for EU cross-references, Commentary for academic analysis. The final memo cites every source with proper legal citations.
Every step is persisted. Sub-queries, individual database results, reasoning chains, source evaluations. Expand any step in the run log to see exactly what the agent searched, what it found, and why it drew a particular conclusion. Full transparency for compliance review or to verify the agent’s work before relying on it.
- Automatic sub-query formulation
- Parallel search across 6 databases
- Cross-jurisdictional analysis (CH + EU)
- Structured research memo output
- Full citation chain with article numbers
- Every step logged: queries, results, reasoning
- Insert findings directly into documents
Read. Research. Edit. Comment. All in one turn.
Sphere’s AI agent doesn’t just execute commands. It orchestrates complex workflows. Ask it to “review the liability section and suggest improvements based on Swiss law” and it will read the document, search legal databases, find relevant clauses, apply edits with tracked changes, and add comments citing its legal sources. Seven tools. One instruction.
The agent asks before it assumes
When DocIQ Sphere encounters ambiguity, missing information, or a decision that requires your judgment, it pauses and presents a structured form. Radio options, text fields, pre-populated defaults, and contextual help text. You answer, hit submit, and the agent resumes exactly where it left off.
A non-compete clause without geographic scope? Sphere asks you to choose before editing. Template variables that need context? Sphere presents a form with all placeholders. Conflicting clauses? Sphere shows the options and lets you decide.
This works everywhere: mid-conversation, inside playbooks, during document creation. The agent never guesses when it can ask.
- Structured forms: radio options + free text
- Contextual help text with legal references
- Pre-populated defaults from document context
- Agent pauses and resumes on submit
- Works in chat, playbooks, and drafting
- Cancel to let the agent proceed without input
Your data. Your infrastructure. Your rules.
Sphere runs where your data policy demands. From managed cloud in CH/EU with professional secrecy guarantees to fully on-premise with zero external dependencies.
Managed Cloud
DefaultHosted on professional-secrecy-grade infrastructure in Switzerland and the EU. Data sovereignty guaranteed under Swiss FADP and GDPR. SOC 2 Type II. Encryption at rest and in transit.
- CH/EU data residency
- Professional secrecy compliance
- Automatic updates & scaling
On-Premise
EnterpriseDeploy Sphere on your own infrastructure. Docker containers on your servers, your network, your security perimeter. Full functionality, zero external data transfer.
- Your servers, your network
- Full feature parity
- Custom SSO / LDAP integration
Air-Gapped
EnterpriseCompletely isolated deployment with local AI models. No internet connection. No telemetry. No external API calls. For organizations where even the metadata must stay internal.
- Zero network dependencies
- Local AI model inference
- Classified environment ready
Every operation is an API call
Full REST API. Native MCP server. Webhooks for every event. Sphere is a platform, not just a UI. Automate document workflows at any scale.
Full REST API
Every Sphere operation is available via REST. Upload documents, trigger playbook runs, apply tracked changes, search your clause bank, and manage data rooms. All programmatically. OpenAPI spec included.
MCP Server
Native Model Context Protocol server turns Sphere into a tool for any MCP-compatible AI agent. Claude, GPT, or your own models can read documents, run reviews, and apply edits through a standardized interface.
Webhooks & Events
Subscribe to document events: batch committed, playbook completed, data room accessed. Push notifications to Slack, trigger downstream pipelines, or feed into your audit system.
Batch Automation
Process hundreds of documents unattended. Queue playbook runs across an entire matter, apply template variables in bulk, or run compliance checks on a document set. Results stream back via webhooks or polling.
# Create a matter and kick off an agent run
curl -X POST https://api.yourinstance.com/v1/agent/run \
-H "Authorization: Bearer $TOKEN" \
-d '{
"prompt": "Run NDA compliance playbook, flag missing clauses",
"documentIds": ["doc_abc123"],
"matterName": "Acme NDA Review"
}'
# => { "matterId": "k57a...", "runId": "ext-m3k..." }
# Resume with a follow-up instruction
curl -X POST .../v1/matters/k57a.../runs \
-H "Authorization: Bearer $TOKEN" \
-d '{ "prompt": "Cap liability at CHF 500k, apply as tracked changes" }'
# Same operations via MCP (Claude, GPT, custom agents)
{ "tool": "run_dociq_agent",
"arguments": { "prompt": "...", "matterId": "k57a..." } }Corporate Legal & RegTech
Review vendor agreements, manage compliance obligations, extract terms at scale. Auto-Playbooks ensure consistent analysis across thousands of documents without manual intervention.
Law Firms & Notaries
Handle cross-border M&A, commercial contracts, regulatory compliance. Playbooks encode your expertise. Data rooms enable secure counterparty collaboration.
Courts & Public Administration
Process judicial correspondence, extract key provisions from submissions, ensure procedural compliance. Multi-format support handles any document type.
Compliance & Risk Officers
GDPR gap analysis, policy reviews, regulatory tracking. Bulk extraction flags issues across entire document libraries. Defensible processes with complete audit trails.
Procurement & Operations
Standardize supplier agreements, extract payment terms and SLAs, flag deviations from approved templates. Clause bank ensures consistency across all contracts.
Any Document-Heavy Team
Sphere works with any document type, not just contracts. Policies, reports, correspondence, regulations. If it’s in Word, PDF, or Excel, Sphere can read, analyze, and edit it.
How DocIQ Sphere compares
Most tools treat documents as text. Sphere treats them as structured legal objects.
Swipe to see all competitors →
Frequently Asked Questions
Ready to modernize your document workflow?
Join the waitlist for early access to DocIQ Sphere. AI-powered document intelligence with Word-native tracked changes, built for teams that take precision seriously.
or register for launch notification
One email when Sphere launches. No spam.