# case.dev - Complete Reference
> The complete API platform for legal technology developers.
This is the extended reference for case.dev. For a concise overview, see [llms.txt](https://case.dev/llms.txt).
## Platform Overview
case.dev is the first complete API platform purpose-built for legal technology developers. It provides a unified suite of AI-powered services and infrastructure tools that enable developers to build production-grade legal applications without managing complex infrastructure.
### The Problem We Solve
Legal tech developers currently face:
- Integration of 5-8 different vendors for document processing, AI, storage, and search
- Vector databases like Pinecone costing $0.23/GB/month - prohibitively expensive for document-heavy applications
- Generic AI APIs lacking legal domain expertise
- Complex vendor management increasing costs and slowing development
### Our Solution
- **One Unified API** - All services through a single platform
- **90% Cost Savings** - Vector storage at 10 credits/GB/month (~$0.10) vs $0.23/GB at competitors
- **Legal-Optimized** - 800+ expert skills, legal embeddings, domain expertise
- **Production-Ready** - Multi-tenant architecture, encryption, audit logs, compliance
## Authentication
All API requests require an API key in the Authorization header:
```
Authorization: Bearer case_sk_live_YOUR_API_KEY
```
Key formats:
- Live keys: `case_sk_live_*`
- Test keys: `case_sk_test_*`
Get your API key at https://console.case.dev
---
## Legal Agent Skills (MCP)
Access 800+ curated legal skills via Model Context Protocol.
**MCP Endpoint:** `https://skills.case.dev/api/mcp`
**Available Tools:**
- `resolve_skill` - Search skills by topic (e.g., "appellate brief", "contract review")
- `read_skill` - Get full skill content by slug
**Browse Skills:** https://agentskills.legal/skills
**Documentation:** https://docs.case.dev/agent-skills
Skills cover: litigation, transactional, corporate, regulatory, research, drafting, analysis.
**Integration Examples:**
Cursor MCP Config:
```json
{
"mcpServers": {
"legal-skills": {
"url": "https://skills.case.dev/api/mcp"
}
}
}
```
Claude Desktop Config:
```json
{
"mcpServers": {
"legal-skills": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://skills.case.dev/api/mcp"]
}
}
}
```
---
## Pricing Model
### Pass-Through Billing
All API usage is billed at upstream provider cost with no markup. You pay exactly what the underlying provider charges.
### Platform Tiers
| Tier | Monthly Price | Description |
|------|---------------|-------------|
| **OSS Pay-Go** | $0 + usage | Pass-through pricing at upstream provider cost. $10 free usage included. 100 RPM LLM rate limit. |
| **Enterprise** | Custom | Custom pricing, BAA/HIPAA available, dedicated support, SLAs. Contact sales. |
### Pass-Through Pricing Reference
| Service | Unit | Price (USD) | Provider |
|---------|------|-------------|----------|
| **Transcription** | 1 hour | $0.30 | AssemblyAI |
| **Text-to-Speech** | 1K characters | $0.12 | ElevenLabs |
| **OCR** | 1K pages | $1.50 | Self-hosted (doctr) |
| **Web Search** | 1K searches | $5.00 | Exa |
| **Translation** | 1M characters | $20.00 | Google Cloud |
| **LLM (CaseMark)** | 1M tokens (input) | $1.50 | Baseten |
| **LLM (CaseMark)** | 1M tokens (output) | $3.50 | Baseten |
| **LLM (3rd party)** | varies | Provider rate | AI Gateway |
| **GPU (H100)** | 1 hour | $3.95 | Modal |
| **DB Compute** | 1 CU-hour | $0.222 | Neon |
| **DB Storage** | 1 GB/month | $0.35 | Neon |
| **Memory** | operation | $0.001 | case.dev |
---
## Service Reference
### 1. Vault API - Document Storage & Semantic Search
Store legal documents with encryption, automatic OCR, semantic chunking, and vector embeddings.
**Endpoints:**
```
POST /vault Create vault
GET /vault/:id Get vault details
PATCH /vault/:id Update vault
DELETE /vault/:id Delete vault
GET /vault/:id/objects List objects
POST /vault/:id/upload Get presigned upload URL
POST /vault/:id/ingest/:objectId Trigger OCR + embedding
POST /vault/:id/search Semantic search
GET /vault/:id/objects/:objectId Get object metadata
GET /vault/:id/objects/:objectId/download Download file
GET /vault/:id/objects/:objectId/text Get extracted text
```
**Create Vault Request:**
```json
{
"name": "Client Documents",
"description": "All client case files",
"default_chunk_size": 512,
"default_chunk_overlap": 50,
"enable_graphrag": true
}
```
**Search Request:**
```json
{
"query": "non-compete clause",
"limit": 10,
"metadata_filter": {
"document_type": "contract"
}
}
```
**Pricing:**
| Service | Credits | Unit |
|---------|---------|------|
| Vault Storage | 3 | per GB/month |
| Vector Storage | 10 | per GB/month |
| Semantic Search | 10 | per 1,000 queries |
First 5GB free on all paid plans.
---
### 2. OCR/Vision API - Document Text Extraction
Extract text from scanned PDFs and images with 95-99% accuracy on clean printed text.
**Endpoints:**
```
POST /ocr/v1/process Submit document for OCR
GET /ocr/v1/:id Check OCR status
GET /ocr/v1/:id/download/text Download plain text
GET /ocr/v1/:id/download/json Download structured JSON
GET /ocr/v1/:id/download/pdf Download searchable PDF
```
**Request:**
```json
{
"file_url": "https://example.com/document.pdf",
"engine": "doctr",
"output_formats": ["text", "json", "pdf"],
"extract_tables": true,
"quality": "legal_grade"
}
```
**Response (JSON output):**
```json
{
"pages": [
{
"page_number": 1,
"words": [
{
"text": "DEPOSITION",
"confidence": 0.98,
"bbox": [100, 150, 400, 200]
}
],
"tables": [{"bbox": [50, 500, 2500, 1500], "csv_url": "..."}]
}
],
"full_text": "DEPOSITION OF JANE DOE...",
"page_count": 150,
"credits_used": 150
}
```
**Supported Formats:** PDF, PNG, JPG, TIFF, BMP (up to 500MB)
**Pricing:**
| OCR Type | Credits/Page | Best For |
|----------|--------------|----------|
| Simple | 1 | Clean PDFs, typed documents |
| Complex | 2 | Tables, forms, mixed layouts |
| Legal Grade | 3 | Depositions, court filings, handwriting |
---
### 3. Voice API - Transcription & Text-to-Speech
Convert audio/video recordings to text with speaker identification and PII redaction.
**Transcription Endpoints:**
```
POST /voice/transcription Submit audio for transcription
GET /voice/transcription/:id Get transcript and status
```
**TTS Endpoints:**
```
POST /voice/v1/speak Generate audio file
POST /voice/v1/speak/stream Stream audio
GET /voice/v1/voices List available voices
```
**Transcription Request:**
```json
{
"audio_url": "https://example.com/deposition.mp3",
"quality": "premium",
"speaker_labels": true,
"redact_pii": true,
"audio_redaction": true,
"auto_chapters": true,
"sentiment_analysis": true,
"word_boost": ["voir dire", "affidavit", "deposition"],
"language_code": "en"
}
```
**Transcription Response:**
```json
{
"id": "trans_abc123",
"status": "completed",
"text": "Speaker A: Please state your name...",
"utterances": [
{
"speaker": "A",
"text": "Please state your name for the record.",
"start": 1.2,
"end": 3.5,
"confidence": 0.96
}
],
"chapters": [
{"summary": "Opening statements", "start": 0, "end": 120}
],
"duration_seconds": 7200,
"credits_used": 130
}
```
**Supported Audio Formats:** MP3, M4A, WAV, FLAC, MP4, MOV, WebM (up to 5GB, 10 hours)
**Pricing:**
| Tier | Credits/Hour | Features |
|------|--------------|----------|
| Standard | 15 | Basic transcription, 90%+ accuracy |
| High Quality | 45 | Speaker labels, timestamps, 95%+ accuracy |
| Premium | 65 | + PII redaction, audio redaction, sentiment |
Per-minute: Standard (0.25), High Quality (0.75), Premium (~1.1)
---
### 4. LLM API - Unified AI Model Access (OpenAI Compatible)
Access 140+ AI models through one OpenAI-compatible API.
**Endpoints:**
```
POST /llm/v1/chat/completions Chat with AI models
POST /llm/v1/embeddings Generate embeddings
GET /llm/v1/models List models with pricing
```
**Chat Request (OpenAI format):**
```json
{
"model": "casemark/core-1",
"messages": [
{"role": "system", "content": "You are a legal research assistant."},
{"role": "user", "content": "Summarize this deposition transcript..."}
],
"max_tokens": 1000,
"temperature": 0.7,
"stream": false
}
```
**Response with Credit Tracking:**
```json
{
"id": "chatcmpl-abc123",
"model": "casemark/core-1",
"choices": [
{
"message": {
"role": "assistant",
"content": "This deposition covers..."
},
"finish_reason": "stop"
}
],
"usage": {
"prompt_tokens": 15234,
"completion_tokens": 487,
"total_tokens": 15721
},
"cost": {
"market_cost": 0.053127,
"actual_cost": 0.042502,
"currency": "USD"
}
}
```
**Available Models by Tier:**
| Tier | Credits/1M Tokens | Example Models |
|------|-------------------|----------------|
| Premium | 2,500 | Claude Opus 4.5, GPT-4 Turbo |
| Standard | 750 | Claude Sonnet 4.5, GPT-4o, CaseMark Core 1 |
| Efficient | 250 | Claude Haiku 4.5, GPT-4o-mini |
| Budget | 35 | DeepSeek V3, Gemini Flash |
**Featured Conversational Models:**
| Model | Provider | Context | Input Price/Token | Output Price/Token | Best For |
|-------|----------|---------|-------------------|-------------------|----------|
| **CaseMark Core 1** | case.dev | 200K | $0.000003 | $0.000015 | Legal workflows (flagship) |
| Claude Sonnet 4.5 | Anthropic | 200K | $0.000003 | $0.000015 | Balanced quality/cost |
| Claude Haiku 4.5 | Anthropic | 200K | $0.00000025 | $0.00000125 | Fast, cost-effective |
| Claude Opus 4.5 | Anthropic | 200K | $0.000015 | $0.000075 | Maximum capability |
| GPT-5 | OpenAI | 128K | $0.000005 | $0.000015 | General purpose |
| GPT-5-mini | OpenAI | 128K | $0.0000001 | $0.0000004 | Ultra-affordable |
| o4-mini | OpenAI | 128K | $0.0000035 | $0.000014 | Reasoning tasks |
| Gemini 2.5 Flash | Google | 1M | $0.000000075 | $0.0000003 | Ultra-fast |
| Gemini 2.5 Pro | Google | 2M | $0.000001 | $0.000004 | Complex reasoning |
| DeepSeek V3 | DeepSeek | 64K | $0.00000014 | $0.00000028 | Open, affordable |
| DeepSeek R1 | DeepSeek | 64K | $0.00000055 | $0.0000022 | Reasoning traces |
| Grok 4 | xAI | 128K | $0.000002 | $0.00001 | Real-time data |
| Qwen 3 32B | Alibaba | 32K | $0.0001/1M | $0.0001/1M | Ultra-cheap |
**Embedding Models:**
| Model | Dimensions | Price/1K Tokens | Best For |
|-------|------------|-----------------|----------|
| text-embedding-3-small | 1536 | $0.00002 | General embeddings |
| text-embedding-3-large | 3072 | $0.00013 | High-quality embeddings |
| voyage-law-2 | 1024 | $0.00012 | **Legal documents** |
| voyage-3.5 | 1536 | $0.00008 | General purpose |
**Features:**
- OpenAI API compatibility (drop-in replacement)
- BYOK (Bring Your Own Key) support
- Automatic cost tracking in every response
- Prompt caching for cost reduction
- Streaming support (SSE)
- Tool use / function calling
---
### 5. Workflows API - Multi-Step Automation
Orchestrate document processing pipelines with durable execution.
**Endpoints:**
```
POST /workflows/v1 Create and execute workflow
GET /workflows/v1/:id Get workflow status
GET /workflows/v1/:id/output Get workflow output
```
**Workflow Definition:**
```json
{
"name": "deposition_analysis",
"steps": [
{
"id": "transcribe",
"service": "voice",
"action": "transcription",
"input": {
"audio_url": "{{input.audio_url}}",
"speaker_labels": true,
"redact_pii": true
}
},
{
"id": "summarize",
"service": "llm",
"action": "chat",
"input": {
"model": "casemark/core-1",
"messages": [
{"role": "user", "content": "Summarize: {{transcribe.text}}"}
]
},
"depends_on": ["transcribe"]
},
{
"id": "store",
"service": "vault",
"action": "upload",
"input": {
"vault_id": "{{input.vault_id}}",
"content": "{{summarize.content}}",
"metadata": {"type": "deposition_summary"}
},
"depends_on": ["summarize"]
}
]
}
```
**Features:**
- Durable execution (survives failures)
- Automatic retries with exponential backoff
- Step-level parallelization
- Conditional execution
- Webhook notifications
**Pricing:**
| Complexity | Credits/Execution | Description |
|------------|-------------------|-------------|
| Simple | 1 | Single API call orchestration |
| Standard | 3 | 2-5 API calls, basic logic |
| Complex | 8 | 5-10 API calls, conditionals |
| Enterprise | 15 | 10+ calls, parallel execution |
Plus underlying service costs (LLM, OCR, etc.)
---
### 6. Actions API - Event-Driven Automation
Execute workflows in response to events.
**Endpoints:**
```
POST /actions/v1 Create action
GET /actions/v1 List actions
DELETE /actions/v1/:id Delete action
```
**Action Definition:**
```json
{
"name": "new_document_processor",
"trigger": {
"type": "vault.object.created",
"vault_id": "vault_abc123"
},
"workflow_id": "wf_xyz789",
"enabled": true
}
```
**Trigger Types:**
- `vault.object.created` - New document uploaded
- `vault.object.ingested` - Document processed
- `schedule` - Cron expression
- `webhook` - External HTTP call
**Pricing:** Free (underlying workflow/service costs apply)
---
### 7. Convert API - FTR Conversion
Convert FTR court recordings to standard M4A audio.
**Endpoint:**
```
POST /convert/v1/process
GET /convert/v1/:id
```
**Request:**
```json
{
"file_url": "https://example.com/recording.ftr",
"output_format": "m4a"
}
```
**Pricing:** 15 credits/hour of audio
---
### 8. Format API - Document Generation
Transform Markdown or JSON into professional documents.
**Endpoint:**
```
POST /format/v1/document
```
**Request:**
```json
{
"content": "# Motion to Dismiss\n\n## Background\n\nPlaintiff filed...",
"format": "markdown",
"output_format": "pdf",
"template": "pleading",
"style": {
"page": {"size": "letter", "margin_top": "1in"},
"typography": {"font_family": "Times New Roman", "font_size": "12pt"}
}
}
```
**Templates:** `pleading`, `letter`, `memo`, `brief`, `contract`, `report`
**Output Formats:** PDF, DOCX, HTML
**Pricing:** 1 credit/document
---
### 9. Email API - Transactional Email
**Endpoint:**
```
POST /email/v1/send
```
**Request:**
```json
{
"from": {"email": "notifications@yourfirm.com", "name": "Smith & Associates"},
"to": [{"email": "client@example.com", "name": "John Client"}],
"subject": "Deposition Transcript Available",
"html": "Your transcript is ready...
",
"attachments": [{"filename": "transcript.pdf", "content_url": "https://..."}]
}
```
**Pricing:** 1 credit per 100 emails
---
### 10. Graph API - Knowledge Graphs (GraphRAG)
Extract entities and relationships from documents.
**Endpoints:**
```
POST /vault/:id/graphrag/init Initialize graph
GET /vault/:id/graphrag/stats Graph statistics
POST /graph/v1/query Query graph
```
**Query:**
```json
{
"vault_id": "vault_abc123",
"query": {
"type": "find_path",
"from_entity": "John Smith",
"to_entity": "Acme Corp",
"max_hops": 3
}
}
```
**Pricing:**
| Service | Credits | Unit |
|---------|---------|------|
| Build Graph | 15 | per 1,000 entities |
| Query Graph | 1 | per query |
First 100 queries/month free.
---
### 11. Billing API
**Endpoints:**
```
GET /billing/v1/usage Current usage summary
GET /billing/v1/history Past invoices
```
**Usage Response:**
```json
{
"period": {"start": "2024-01-01T00:00:00Z", "end": "2024-01-31T23:59:59Z"},
"usage": {
"total_cost_usd": 8.47,
"free_allowance_usd": 10.00,
"free_allowance_remaining_usd": 1.53,
"has_payment_method": false
},
"by_service": {
"llm": {"cost_usd": 4.50, "units": "600K tokens"},
"transcription": {"cost_usd": 1.95, "units": "3 hours"},
"ocr": {"cost_usd": 1.50, "units": "1,000 pages"},
"storage": {"cost_usd": 0.52, "units": "17 GB"}
}
}
```
---
### 12. Orbit Compute - Application Hosting
Deploy web applications and GPU workloads.
**Endpoints:**
```
POST /deployments/v1 Create deployment
GET /deployments/v1 List deployments
POST /deployments/v1/:id/pause Pause deployment
POST /deployments/v1/:id/resume Resume deployment
GET /deployments/v1/:id/logs Get logs
```
**Deployment Request:**
```json
{
"name": "legal-portal",
"repository": "https://github.com/firm/legal-portal",
"branch": "main",
"framework": "nextjs",
"environment_variables": {
"DATABASE_URL": "postgresql://...",
"API_KEY": "case_sk_live_..."
},
"instance_size": "medium"
}
```
**Features:**
- One-click deployment from GitHub
- Live URLs on case.systems domain
- Container and static site support
- Pause/resume for cost savings
- Real-time deployment logs
- Per-second billing
**Instance Pricing (billed separately, not credits):**
| Size | vCPU | Memory | Price/Hour |
|------|------|--------|------------|
| Small | 1 | 2GB | $0.03 |
| Medium | 2 | 4GB | $0.08 |
| Large | 4 | 8GB | $0.16 |
**GPU Pricing (pass-through from Modal):**
| GPU | VRAM | Price/Hour |
|-----|------|------------|
| T4 | 16GB | $0.59 |
| L4 | 24GB | $0.83 |
| A100 | 80GB | $2.50 |
| H100 | 80GB | $3.95 |
| B200 | 192GB | Custom |
---
### 13. Payments Platform (Early Access)
IOLTA-compliant trust accounting and payment infrastructure.
**Features:**
- Double-entry ledger for all money movements
- Trust account management
- Holds and approval workflows
- Stripe Connect integration
- Settlement processing
- Complete audit trails
**Status:** Early access for select partners. Contact sales@case.dev for access.
---
### 14. Translation API - Multilingual Document Processing
Neural machine translation powered by Google Cloud Translation, supporting 100+ languages.
**Endpoints:**
```
POST /translate/v1/translate Translate text
POST /translate/v1/detect Detect language
GET /translate/v1/languages List supported languages
```
**Translate Request:**
```json
{
"text": "This agreement shall be governed by the laws of California.",
"target_language": "es",
"source_language": "en"
}
```
**Response:**
```json
{
"translated_text": "Este acuerdo se regirá por las leyes de California.",
"detected_language": "en",
"confidence": 0.99
}
```
**Features:**
- Auto language detection
- Batch translation (up to 100 texts)
- HTML preservation
- 100+ language pairs
**Pricing:** 3 credits per 1,000 characters
---
### 15. Web Search API - Search & AI Answers
Programmatic web search with AI-powered answer generation.
**Endpoints:**
```
POST /search/v1/search Web search with domain filtering
POST /search/v1/answer Get AI answers with citations
POST /search/v1/research Deep multi-step research
```
**Answer Request:**
```json
{
"query": "What is the statute of limitations for medical malpractice in California?",
"includeDomains": ["law.cornell.edu", "findlaw.com"]
}
```
**Response:**
```json
{
"answer": "In California, the statute of limitations for medical malpractice is 3 years from the date of injury or 1 year from discovery [1]...",
"citations": [
{"title": "California Code of Civil Procedure § 340.5", "url": "..."}
]
}
```
**Features:**
- Domain filtering (include/exclude)
- AI-generated answers with citations
- Multi-step research for comprehensive reports
**Note:** Web data only - use specialized platforms like Lexis or Westlaw for case law/statutes.
---
### 16. Databases API - Serverless PostgreSQL
Serverless PostgreSQL databases powered by Neon with instant branching.
**Endpoints:**
```
POST /database/v1/projects Create database project
GET /database/v1/projects List projects
GET /database/v1/projects/:id Get project details
GET /database/v1/projects/:id/connection Get connection string
POST /database/v1/projects/:id/branches Create branch
```
**Create Project Request:**
```json
{
"name": "case-management-db",
"region": "aws-us-east-1"
}
```
**Response:**
```json
{
"id": "proj_abc123",
"name": "case-management-db",
"region": "aws-us-east-1",
"connection_uri": "postgresql://...",
"status": "active"
}
```
**Features:**
- Serverless auto-scaling
- Instant branching (dev/staging/prod copies)
- Connection pooling (PgBouncer included)
- PostgreSQL 17 with full feature support
- Encrypted at rest and in transit
**Regions:** aws-us-east-1 (default), aws-us-east-2, aws-us-west-2, aws-eu-central-1, aws-eu-west-1, aws-eu-west-2, aws-ap-southeast-1, aws-ap-southeast-2
---
### 17. Memory API - Persistent Agent Memory
Persistent, searchable memory for AI agents across conversations.
**Endpoints:**
```
POST /memory/v1 Add memories from conversation
POST /memory/v1/search Search memories semantically
GET /memory/v1 List memories
DELETE /memory/v1/:id Delete memory
POST /memory/v1/delete-many Bulk delete by tags
```
**Add Memory Request:**
```json
{
"messages": [
{"role": "user", "content": "Remember that this client prefers conservative investments and is risk-averse."}
],
"tag_1": "client_123",
"tag_2": "matter_456",
"category": "preference"
}
```
**Response:**
```json
{
"results": [
{"id": "mem_abc", "memory": "Client prefers conservative investments", "event": "ADD"},
{"id": "mem_def", "memory": "Client is risk-averse", "event": "ADD"}
]
}
```
**Search Request:**
```json
{
"query": "investment preferences",
"tag_1": "client_123",
"top_k": 5
}
```
**Features:**
- Automatic fact extraction from conversations
- Semantic search (find by meaning, not keywords)
- 12 indexed tag fields for filtering
- Deduplication and conflict resolution
- Custom extraction prompts for domain-specific needs
**Memory Events:**
- `ADD` - New memory created
- `UPDATE` - Existing memory updated/merged
- `DELETE` - Contradicting info caused deletion
- `NONE` - Duplicate, already exists
**Pricing:** $0.001 per operation (add, search). List/get/delete are free.
---
### 18. SuperDoc API - Document Conversion
Convert documents between formats and populate DOCX templates.
**Endpoints:**
```
POST /superdoc/v1/convert Convert between formats
POST /superdoc/v1/annotate Populate DOCX template fields
```
**Convert Request:**
```json
{
"file": "",
"from": "md",
"to": "pdf"
}
```
**Annotate Request (Template Population):**
```json
{
"document": {
"url": "https://your-bucket.s3.amazonaws.com/templates/contract.docx"
},
"fields": [
{"id": "client_name", "type": "text", "value": "Acme Corporation"},
{"id": "contract_date", "type": "date", "value": "2025-01-15"},
{"id": "contract_amount", "type": "number", "value": 50000}
],
"output_format": "pdf"
}
```
**Supported Conversions:**
| From | To | Use Case |
|------|-----|----------|
| docx | pdf | Finalize Word documents |
| md | docx | Generate editable Word docs |
| md | pdf | Create PDFs from Markdown |
| html | docx | Convert web content |
**Features:**
- High-fidelity formatting preservation
- Template field population
- Multiple input methods (upload, URL, base64)
---
### 19. AI Governance - Provider Control
Control which AI providers handle your organization's requests.
**Endpoints:**
```
GET /governance List policies
POST /governance Create policy
PUT /governance/:id Update policy
DELETE /governance/:id Delete policy
GET /governance/audit Query audit logs
```
**Create Policy Request:**
```json
{
"name": "SOC 2 Compliant",
"rules": {
"minimumTrustTier": "most_trusted",
"requireSoc2": true,
"enforcement": "hard_block"
}
}
```
**Use in LLM Request:**
```json
{
"model": "claude-sonnet-4-20250514",
"messages": [{"role": "user", "content": "Analyze this contract"}],
"governance_policy": 2
}
```
**Trust Tiers:**
| Tier | Description | Use Case |
|------|-------------|----------|
| `most_trusted` | Full enterprise compliance (SOC 2, HIPAA, BAA, ZDR) | Healthcare, financial |
| `trusted` | Good compliance, may lack some certifications | General enterprise |
| `sketchy` | Limited compliance info | Development/testing only |
| `untrustworthy` | Subject to national security laws | Always blocked |
**Compliance Certifications:** SOC 2 Type II, HIPAA, BAA, ZDR (Zero Data Retention), ISO 27001, GDPR
**Enforcement Modes:**
- `hard_block` - Request fails with 403
- `soft_block` - Request fails, violation logged
- `warn` - Request proceeds, violation logged
**Features:**
- Trust tier enforcement
- Compliance certification requirements
- Complete audit logging
- Per-request policy selection
---
## Error Handling
All errors follow this format:
```json
{
"error": {
"code": "INVALID_API_KEY",
"message": "The API key provided is invalid",
"type": "authentication_error"
}
}
```
Common error codes:
- `401` - Invalid or missing API key
- `403` - Insufficient permissions or credits exhausted
- `404` - Resource not found
- `429` - Rate limit exceeded (check `Retry-After` header)
- `500` - Internal server error
---
## Rate Limits
Rate limits vary by endpoint and plan. Headers returned:
- `X-RateLimit-Limit`: Requests allowed per window
- `X-RateLimit-Remaining`: Requests remaining
- `X-RateLimit-Reset`: Unix timestamp when limit resets
- `X-Credits-Remaining`: Credits remaining in current period
---
## SDKs & Libraries
**Python:**
```bash
pip install casemark
```
```python
from casemark import CaseMark
client = CaseMark(api_key="case_sk_live_...")
response = client.llm.chat.completions.create(
model="casemark/core-1",
messages=[{"role": "user", "content": "Hello"}]
)
print(f"Credits used: {response.usage.credits_used}")
```
**TypeScript/JavaScript:**
```bash
npm install @casemark/sdk
```
```typescript
import { CaseMark } from '@casemark/sdk';
const client = new CaseMark({ apiKey: 'case_sk_live_...' });
const response = await client.llm.chat.completions.create({
model: 'casemark/core-1',
messages: [{ role: 'user', content: 'Hello' }]
});
console.log(`Credits used: ${response.usage.credits_used}`);
```
**Additional SDKs:**
- PHP (available)
- C# (available)
- Java (available)
---
## Usage Scenarios
### Scenario 1: Small Law Firm (OSS Pay-Go)
| Activity | Volume | Cost (USD) |
|----------|--------|------------|
| OCR | 50 pages | $0.08 |
| Transcription | 5 hours | $1.50 |
| LLM (CaseMark Core) | 50K tokens | $0.18 |
| DB Storage | 10 GB | $3.50 |
| Web Search | 500 queries | $2.50 |
| **Total** | | **$7.76** (within $10 free allowance) |
### Scenario 2: Solo Practitioner (OSS Pay-Go)
| Activity | Volume | Cost (USD) |
|----------|--------|------------|
| OCR | 200 pages | $0.30 |
| Transcription | 10 hours | $3.00 |
| LLM (CaseMark Core) | 200K tokens | $0.70 |
| DB Storage | 25 GB | $8.75 |
| Translation | 50K chars | $1.00 |
| **Total** | | **$13.75** ($3.75 after $10 free allowance) |
### Scenario 3: Mid-Size Firm (OSS Pay-Go)
| Activity | Volume | Cost (USD) |
|----------|--------|------------|
| OCR | 1,000 pages | $1.50 |
| Transcription | 20 hours | $6.00 |
| LLM (CaseMark Core) | 500K tokens | $1.75 |
| DB Compute | 100 CU-hours | $22.20 |
| DB Storage | 100 GB | $35.00 |
| **Total** | | **$66.45** (pass-through, no markup) |
---
## Common Integration Patterns
### RAG (Retrieval-Augmented Generation)
1. Upload documents to Vault
2. Ingest to create embeddings
3. Search for relevant chunks
4. Pass chunks to LLM with user query
### Document Processing Pipeline
1. Upload PDF to Vault
2. OCR if scanned
3. Extract text and metadata
4. Generate summary with LLM
5. Store results back in Vault
### Deposition Analysis
1. Convert FTR to M4A (if needed)
2. Transcribe with speaker labels
3. Store transcript in Vault
4. Extract entities with GraphRAG
5. Generate timeline and summary
---
## Links
- **Documentation**: https://docs.case.dev
- **API Reference**: https://docs.case.dev/api-reference
- **Console**: https://console.case.dev
- **Pricing**: https://case.dev/pricing
- **Status**: https://status.case.dev
- **Security**: https://trust.casemark.com
- **Support**: support@case.dev
- **Sales**: sales@case.dev
---
## Related Products
- **Legal Agent Skills** (https://agentskills.legal): 800+ curated legal prompts, templates, and methodologies accessible via MCP
- **MCP Server** (https://docs.case.dev/mcp): Integration with AI coding assistants like Cursor and Claude Desktop
- **Console** (https://console.case.dev): Web dashboard for API management