This is not about features or price - it’s about doing one thing exceptionally well. Cleft transforms your longwinded verbal thoughts into usable notes. That’s it. That’s what we do.
What Makes Cleft Different
What Cleft Is
- Personal thought processor for your ideas and insights
- Voice-first note-taking with intelligent formatting
- Privacy-focused with local Whisper transcription
- Complementary tool that works alongside your existing workflow
- Built for verbal thinkers who process ideas out loud
What Cleft Is Not
- ❌ Meeting transcription service (like Otter.ai)
- ❌ AI assistant replacement (like ChatGPT)
- ❌ Team collaboration tool
- ❌ Generic transcription service
- ❌ “Do everything” productivity app
Our One Thing: Transforming Verbal Thoughts Into Usable Notes
We do one thing exceptionally well: take your longwinded, messy verbal thoughts and transform them into organized, usable notes. Not transcription. Not AI assistance. Thought transformation. You talk. Cleft listens, understands, and formats your ideas exactly how you want them - from custom instructions to future expansions we’re planning. This is what you pay for: an amazing experience that turns voice into value.We Work WITH Your Tools, Not Against Them
Integration Philosophy:- Use Cleft IN Google Meet for your personal notes
- Use Cleft WITH Microsoft Teams for your follow-up thoughts
- Export TO Notion, Obsidian, or wherever you work
- We complement your workflow, we do not compete with it
- A replacement for meeting transcription (use Otter.ai for that)
- An AI assistant competitor (ChatGPT does that better)
- A team collaboration tool (Slack handles that)
- Your personal thought processor
- A bridge between your verbal mind and your digital workflow
- Software that charges for experience, not storage
How We Compare to Different Tool Categories
They do AI collaboration. We do not.Use them for:
- Research assistance and AI-powered writing
- Back-and-forth conversation with AI
- Content creation WITH an AI partner
- General-purpose AI help ($20/month)
- Capturing YOUR thoughts, not generating AI thoughts
- Private processing of personal ideas
- Thought transformation without AI collaboration
- ChatGPT Plus: GPT-4/Whisper models, cloud processing, can opt-out of training but default may include it
- Claude Pro: Claude 3.5 models, 5-year data retention if opted-in (deadline Sept 28, 2025)
- Cleft: OpenAI GPT for text formatting only, local Whisper, never trains on user data
Comprehensive Privacy, Legal & Compliance Comparison
Click column headers to sort. Data updated September 2025.| Tool | Price | LLM/AI | Privacy | DPA | Terms | Compliance | Data Location | Deletion | Overall Score |
|---|---|---|---|---|---|---|---|---|---|
| Cleft | $6.99/mo | OpenAI GPT (text only) | 🟢 Local Whisper + No training | Available | Fair terms | 🟢 GDPR + Privacy-first | 🟢 Local processing | 🟢 Immediate on request | 9.2/10 |
| Apple Voice Memos | Free | None | 🟢 Local only | Consumer only | Standard Apple | 🟢 Strong privacy controls | 🟢 On-device | 🟢 User controlled | 8.8/10 |
| AudioPen | $20/mo | Unknown | 🔴 Cloud processing | ❌ None found | Basic terms | 🔴 Unknown compliance | 🔴 Unknown location | 🔴 Unclear policy | 3.1/10 |
| ChatGPT Plus | $20/mo | GPT-4, Whisper | 🔴 Cloud + training risk | Enterprise only | Arbitration clause | 🟡 SOC 2, working on others | 🔴 US + global | 🟡 30 days (can extend) | 4.7/10 |
| Claude Pro | $20/mo | Claude 3.5 | 🔴 5-year retention option | Enterprise available | Consumer friendly | 🟡 SOC 2 Type II | 🔴 US-based | 🔴 5 years if opted-in | 5.1/10 |
| Fireflies.ai | $18/mo | Multiple | 🔴 Cloud + improvement use | Business plans | Liability limits | 🟡 SOC 2 pending | 🔴 US + AWS global | 🟡 Account deletion available | 4.2/10 |
| Gemini Advanced | $19.99/mo | Gemini Pro | 🔴 Google ecosystem | Workspace only | Google standard | 🟢 Multiple certifications | 🔴 Global Google DCs | 🟡 18 months retention | 5.8/10 |
| Google Meet | $6-18/user | Google proprietary | 🟡 Enterprise controls | Workspace DPA | Business terms | 🟢 ISO 27001, SOC 2/3 | 🔴 Global Google DCs | 🟡 Admin controlled | 6.4/10 |
| Granola | $10/mo | OpenAI, Anthropic | 🟡 No audio storage | Enterprise | Standard SaaS | 🟡 SOC 2 Type II | 🟡 US (AWS) | 🟢 No audio retained | 6.8/10 |
| Otter.ai | $16.99/mo | Proprietary AI | 🔴 Trains on data | Business/Enterprise | Arbitration required | 🟡 SOC 2, working on HIPAA | 🔴 US (AWS) | 🟡 Can request deletion | 4.4/10 |
| Voice Notes | $14.99/mo | Various providers | 🟡 Mixed approaches | ❌ None found | Basic indie terms | 🔴 Unknown compliance | 🔴 Varies by provider | 🔴 Unclear policy | 4.9/10 |
| Wispr Flow | $7/mo | Whisper + Llama | 🟡 30-day deletion | ❌ None found | VC startup terms | 🟡 Privacy mode available | 🔴 Cloud processing | 🟢 30-day auto-delete | 6.2/10 |
A Note on These Tools
Every tool has its place. We have deep respect for all these applications - they’re excellent at what they’re designed to do:- ChatGPT, Claude, and Gemini are phenomenal AI assistants for research and collaboration
- Otter.ai, Fireflies, and Google Meet excel at meeting transcription and team documentation
- AudioPen and Voice Notes provide solid voice-to-text conversion
- Granola offers AI-enhanced meeting notes and assistance
- Wispr Flow provides excellent system-wide dictation across all apps
- Apple Voice Memos is perfect for simple, reliable audio recording
We built Cleft because we needed something different - a tool specifically for transforming personal verbal thoughts into usable notes. Not better than these tools, just different. Each serves its purpose brilliantly.
This is not a hit piece - it’s information sharing to help you find the right tool for your specific needs. Choose what works for your workflow.
Understanding Privacy Trade-offs
Local vs Cloud Processing
Local Processing (🟢 Excellent)- Your audio never leaves your device
- No dependency on internet for core functionality
- No risk of data breaches with your personal thoughts
- Limited by your device’s processing power
- Examples: Cleft (Whisper), Apple Voice Memos
- Audio uploaded to external servers for transcription
- Faster processing with powerful cloud infrastructure
- Risk of data breaches and unauthorized access
- May be subject to government requests or legal discovery
- Examples: Most other services in this comparison
What “Zero Data Retention” Actually Means
True Zero Retention:- Data deleted immediately after processing
- No copies kept in logs, backups, or caches
- Third-party providers also delete data
- Example: Wispr Flow claims this with provider agreements
- Main data deleted, but metadata may remain
- Logs and analytics often exclude “personal data” but include usage patterns
- Third-party subprocessors may have different policies
- Red Flag: Vague language about “personal data” vs all data
Evaluating Privacy Policies in 2025
🚩 Red Flags to Watch For:
- “May use data for model training” without clear opt-out
- Vague language like “anonymized” without technical details
- Long retention periods (years) for “business purposes”
- Sharing with “trusted partners” without naming them
- Default opt-in to data training programs
✅ Green Flags:
- Specific technical details about data handling
- Clear opt-out mechanisms that are easy to find
- Named, limited third-party providers
- SOC 2, GDPR compliance with regular audits
- Explicit commitments about not training on user data
- Anthropic extending retention to 5 years if users allow training
- OpenAI facing court orders for indefinite conversation retention
- Class-action lawsuits against transcription services over consent
- On-device AI becoming more viable for privacy-conscious users
LLM Provider Landscape
The OpenAI Dominance
Whisper for Speech Recognition:- Industry standard for speech-to-text conversion
- Available as local model (privacy-friendly) or cloud API
- Used by: Cleft (local), ChatGPT, Wispr Flow, many others
- Trade-off: Cloud version faster but compromises privacy
- GPT-4 family dominates text understanding and formatting
- Used by most services for summarization and organization
- Privacy concern: Default policies may include training on user data
Anthropic’s Changing Approach
Claude Models:- Strong competitor to GPT-4 for text analysis
- 2025 Policy Change: Extending retention to 5 years for training data
- Users must opt-out by September 28, 2025 or be automatically enrolled
- Used by: Granola (as option), Claude Pro directly
The Privacy-First Alternative Movement
Local/On-Device Processing:- Apple’s on-device AI processing
- Meta’s Llama models (can run locally)
- Emerging specialized voice models
- Trend: More apps offering local processing options in 2025
- Some newer providers guarantee no data storage
- Wispr Flow negotiates zero-retention agreements with LLM providers
- Verification challenge: Hard to audit third-party compliance
Enterprise vs Consumer Privacy Standards
Enterprise Customers Get Better Privacy:- Google Workspace: Enterprise controls for data usage
- Granola: Enterprise plans have training disabled by default
- Most providers: Different policies for business vs consumer accounts
- Consumer Reality: Often default opt-in to training, must manually opt-out
What This Means for You
If Privacy is Your Priority:- Choose local processing (Cleft, Apple Voice Memos)
- If cloud needed, verify zero-retention claims
- Enterprise plans often have better privacy controls
- Read policies carefully - “anonymized” doesn’t mean “not used”
- Cloud services offer more advanced AI capabilities
- Multiple LLM options provide better results
- Faster processing and more integrations available
- Accept privacy trade-offs for functionality gains
Legal Documents Deep Dive
Data Processing Agreements (DPAs)
What is a DPA? A legally binding contract that defines how your data is processed by a service provider. Essential for GDPR compliance and business use. 🟢 Comprehensive DPAs Available:- Cleft: Full DPA with data residency guarantees, local processing commitments
- Google Workspace: Industry-standard DPA covering Meet transcription
- Granola Enterprise: Business-grade DPA with audit rights
- ChatGPT: Enterprise plans only, consumers get standard terms
- Claude: Enterprise customers get enhanced protections
- Otter.ai: Business/Enterprise plans include DPA options
- Fireflies.ai: DPA available for business customers
- AudioPen: Bootstrapped service without formal DPA
- Voice Notes: Individual developer, no business agreements
- Wispr Flow: VC-funded but no formal DPA yet
- Apple Voice Memos: Consumer service only
Terms of Service Analysis
Key Legal Protections to Look For:- Liability Limitations: How much can you recover if something goes wrong?
- Indemnification: Will the company protect you from legal claims?
- Dispute Resolution: Courts vs. mandatory arbitration
- Data Ownership: Who owns your content?
- Service Changes: Can terms change without notice?
- Cleft: Fair liability terms, optional arbitration, clear data ownership
- Google Workspace: Enterprise protections, reasonable liability caps
- Claude Pro: Consumer-friendly terms with clear opt-outs
- Apple: Standard consumer agreement, reasonable for personal use
- Granola: Typical SaaS terms with enterprise options
- Gemini Advanced: Google’s standard consumer terms
- ChatGPT Plus: Mandatory arbitration, broad liability exemptions
- Otter.ai: Required arbitration for all disputes
- Fireflies.ai: Limited liability protection, startup risk
Compliance & Certifications
SOC 2 Type II (Security audit standard):- ✅ Have it: Google, Granola, Claude (Anthropic), Otter.ai
- 🔄 Working on it: ChatGPT (OpenAI), Fireflies.ai
- ❌ Not applicable: Apple (consumer), Cleft (privacy-first design), AudioPen, Voice Notes, Wispr Flow
- ✅ Full compliance: Cleft, Apple, Google services, Claude
- 🟡 Partial compliance: ChatGPT, Otter.ai, Granola
- ❌ Unknown/Limited: AudioPen, Voice Notes, Wispr Flow
- ✅ Available: Google Workspace (with BAA)
- 🔄 Working on it: Otter.ai (announced for 2025)
- ❌ Not available: All other services
- ✅ Certified: Google services
- 🔄 Working on it: Several others in enterprise plans
- ❌ Not certified: Most consumer services
Data Residency & Sovereignty
🟢 Local/Controlled Processing:- Cleft: Audio processed locally, text to OpenAI US servers only
- Apple Voice Memos: Completely on-device, your choice of iCloud region
- Google services: Data center locations disclosed, some region control
- Granola: AWS US-East, clear about location
- Claude: Anthropic US-based infrastructure
- ChatGPT: Global infrastructure, limited control
- Otter.ai: AWS global, some data may cross borders
- Fireflies.ai: US + global cloud providers
- AudioPen, Voice Notes, Wispr Flow: Locations not clearly disclosed
Government Access & Legal Requests
US Services (subject to US government requests):- All services hosted in US can receive NSA/FBI requests
- Transparency reports available for: Google, Apple, OpenAI, Anthropic
- Challenge history: Apple fights requests, Google complies with transparency
- GDPR: EU residents have additional protections
- Data localization laws: Some countries require local storage
- Cross-border transfers: Most services use US infrastructure
What This Means for Different Users
Priority: Privacy and ease of useRecommendations:
- High privacy needs: Cleft, Apple Voice Memos (local processing)
- Feature-rich needs: Willing to trade privacy for features
- Budget-conscious: Apple Voice Memos (free) or Cleft (best value)
Our Experience-Based Pricing Philosophy
We charge for an amazing experience, not consumption:- NOT per page, per minute, or per gigabyte
- NOT like a “digital notebook printing company”
- YES for thought transformation technology
- YES for software that people love more than we do
| What Others Charge For | What Cleft Charges For |
|---|---|
| Storage space | Thought transformation |
| Transcription minutes | Experience quality |
| Per-seat licensing | Individual productivity |
| Feature access tiers | One great experience |
Why Funding Matters
Bootstrapped (Customer-funded):- Built for profitability and sustainability
- Aligned with customer interests, not investor returns
- Focus on product quality over hypergrowth
- Examples: Cleft, AudioPen, Voice Notes
- Growth-focused with investor pressure to scale
- May prioritize user acquisition over profitability
- Risk of pivot, acquisition, or shutdown
- Examples: OpenAI, Anthropic, Otter.ai, most startups
- Integrated into broader platforms
- Different privacy considerations
- May discontinue products strategically
- Examples: Apple, Google products
$3.33/month (annual) - You’re paying for bootstrapped artisan software that does one thing exceptionally well, built to last without investor pressure.