Pricing
Start free. Upgrade when you need deeper analysis.
Plan Comparison
| Feature | Free | Starter $29/mo |
Pro $99/mo |
Enterprise $599/mo |
|---|---|---|---|---|
| Orphans lens | Yes | Yes | Yes | Yes |
| Obsolescence lens | Yes | Yes | Yes | Yes |
| Contradictions lens | - | Yes | Yes | Yes |
| Gaps lens | - | Yes | Yes | Yes |
| OmniQ lenses | - | - | - | Yes (4 lenses) |
| Write-back actions | - | Manual | Manual + Auto Safe | All modes (incl. Auto Full) |
| Smart Query | 500/month | Unlimited | Unlimited | Unlimited |
| Chat | Yes | Yes | Yes | Yes |
| API access | Yes | Yes | Yes | Yes |
| Background scheduler | - | Every 6h | Every 6h | Configurable |
| Perception protocols | Yes | Yes | Yes | Yes |
| Knowledge protocol | Yes | Yes | Yes | Yes |
| Slack notifications | - | Yes | Yes | Yes |
| PDF upload quota | 100 chunks | 5,000 chunks | 50,000 chunks | Unlimited |
| Document Editor | Yes | Yes | Yes | Yes |
| Observation Auditor | Yes | Yes | Yes | Yes |
| Supported providers | Qdrant, Pinecone, Weaviate, ChromaDB (or use PDF upload with no vector DB) | |||
Cost Per Cognitive Cycle
These are approximate LLM costs for a single cognitive cycle on a 1,000-document RAG. You pay your subscription fee plus these usage costs.
| Component | Free | Pro | Enterprise |
|---|---|---|---|
| Orphans + Obsolescence | $0 | $0 | $0 |
| Contradictions (up to 50 Haiku calls) | - | ~$0.05 | ~$0.05 |
| Gaps (1 Haiku call) | - | ~$0.001 | ~$0.001 |
| OmniQ (1-2 Sonnet calls) | - | - | ~$0.02-0.04 |
| Programmer (1 Sonnet call) | ~$0.01 | ~$0.01 | ~$0.01 |
| Cartographer (1 Sonnet call, if delta >15%) | ~$0.01 | ~$0.01 | ~$0.01 |
| Total (approximate) | ~$0.02 | ~$0.07 | ~$0.10-0.13 |
Costs scale with document count. The numbers above assume ~1,000 documents. Contradiction detection is capped at 50 document pairs per cycle to control costs.
What's Included
Free
- Structural analysis (orphans, stale docs)
- Observation Auditor (trust scores)
- PDF upload (100 chunks)
- Document Editor
- Multi-source chat with citations
- 500 Smart Queries per month
- Dashboard + API access
Starter $29/mo
- Everything in Free
- Contradiction detection (LLM-powered)
- Coverage gap analysis
- PDF upload (5,000 chunks)
- Write-back actions (manual mode)
- Unlimited Smart Queries
- Slack notifications
- Auto-scan every 6 hours
Pro $99/mo
- Everything in Starter
- PDF upload (50,000 chunks)
- Auto Safe write-back mode
- Faster scan intervals
- Priority support
Enterprise $599/mo
- Everything in Pro
- OmniQ perception lenses (Monade, Symbiote, Architect, Empath)
- Unlimited PDF upload
- Auto Full write-back mode
- Configurable scan schedule
- Confidence scoring on all observations
Frequently Asked Questions
Can I try Seahorse before paying?
Yes. The Free tier is fully functional with structural analysis (orphans, obsolescence), chat, perception protocols, and 500 Smart Queries per month. No credit card required. You only need to upgrade if you want LLM-powered lenses (contradictions, gaps) or OmniQ perception.
What counts as a cognitive cycle?
One cognitive cycle is a full run of the 8-phase cognitive loop: scanning, analyzing, auditing, evaluating, programming, executing, superchunking, and cartographing. Each cycle can run up to 5 iterations internally. You trigger cycles manually or let the background scheduler run them automatically (Starter+ tiers).
Can I downgrade my plan?
Yes. Tier changes take effect at the next billing period. When you downgrade, you keep access to your current tier's features until the period ends. Your data (observations, protocols, snapshots) is preserved regardless of tier.
Do I need a vector database?
No. You can upload PDFs directly or use the Document Editor without any vector database. If you do have one, Seahorse supports Qdrant, Pinecone, Weaviate, and ChromaDB with full read, write, and change detection. You can use all three input methods simultaneously.
Does Seahorse modify my documents?
Only when you authorize it. In Manual mode (the default), Seahorse finds issues and suggests fixes, but you approve each action individually. Even in automatic modes, every write-back creates a snapshot so you can undo any change with one click.
How are LLM costs billed?
LLM costs (for contradiction detection, gap analysis, OmniQ lenses, and protocol generation) are included in your subscription. The cost breakdown in the table above is for transparency so you understand what each cycle does. You are not billed separately for LLM usage.
Is my data secure?
Yes. Each client gets isolated storage. Authentication uses JWT with bcrypt-hashed passwords. API keys are SHA-256 hashed. All write-back actions create snapshots. URLs are validated against SSRF attacks. See our API Reference for security details.
Ready to get started?
Create your free account and start analyzing documents in under 5 minutes.
Create Free Account