How do I secure Cohere applications?

Integrate Wardstone Guard with your Cohere application to detect prompt injections, jailbreaks, harmful content, and data leakage. Wardstone validates inputs before they reach Cohere and screens outputs before returning them to users.

Cohere + Wardstone

Enterprise RAG Security

Secure your Cohere Command applications with Wardstone Guard. Protect enterprise RAG pipelines and agentic workflows from prompt attacks and data leakage.

14 Supported Models· all protected with sub-30ms latency

Command ACommand A ReasoningCommand A VisionCommand R+Command RCommand R 08-2024+8 more

Try the Playground Visit Cohere

Why Secure Cohere?

RAG Poisoning

High Risk

Malicious documents in your knowledge base can inject harmful instructions through retrieval.

Rerank Manipulation

Medium Risk

Attackers can craft content that games reranking to surface malicious results.

Embedding Space Attacks

Medium Risk

Adversarial inputs can manipulate semantic search to return harmful content.

Security Considerations

1
Cohere's RAG focus means external documents are a major attack surface
2
Embed models don't have safety training - only generation models do
3
Rerank models can surface malicious content to top positions
4
Wardstone protects the full RAG pipeline, not just generation

How to Integrate

Add Wardstone to your pipeline
Install Wardstone SDK alongside the Cohere client.
Validate user queries
Screen search queries before embedding and retrieval.
Scan retrieved documents
Validate retrieved content before passing to the generation model.
Screen generated responses
Check final outputs for data leakage and policy violations.

Pricing Note

Cohere charges separately for embed, rerank, and generation. Wardstone provides unified security across all Cohere APIs.

Secure Cohere with Wardstone

# Step 1: Check user input with Wardstone
curl -X POST "https://api.wardstone.ai/v1/detect" \
  -H "Authorization: Bearer YOUR_WARDSTONE_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text": "User message here"}'
 
# Response: { "prompt_attack": { "detected": false, ... } }
 
# Step 2: If safe, send to Cohere (v2 API)
curl -X POST "https://api.cohere.com/v2/chat" \
  -H "Authorization: Bearer YOUR_COHERE_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "command-a-03-2025",
    "messages": [{"role": "user", "content": "User message here"}]
  }'
 
# Step 3: Check Cohere response with Wardstone before returning to user

Common Use Cases

Enterprise search and retrieval

RAG-powered chatbots

Document Q&A systems

Knowledge management

Semantic search applications

All Supported Cohere Models

Wardstone Guard protects all Cohere models with the same comprehensive security coverage. Whether you're using the latest releases or legacy models still in production, every API call is protected.

Command A

Command A Reasoning

Command A Vision

Command R+

Command R

Command R 08-2024

Command R+ 08-2024

Embed v4.0

Embed v3.0 English

Embed v3.0 Multilingual

Rerank v3.5

Rerank v3.0