Credentials - Flowise

What are Credentials?

Credentials in Flowise are secure storage objects for API keys, tokens, and authentication information required by nodes to connect to external services. They provide centralized, encrypted management of sensitive data, ensuring your integrations work without exposing secrets in your workflow configurations.

Credentials are encrypted before storage and are never included in exported chatflows or agentflows, protecting your sensitive data.

Credential Architecture

Credential Structure

Credentials are stored as database entities:

interface ICredential {
  id: string              // Unique credential ID
  name: string           // User-defined name (e.g., "My OpenAI Key")
  credentialName: string // Type identifier (e.g., "openAIApi")
  encryptedData: string  // Encrypted credential values
  createdDate: Date
  updatedDate: Date
  workspaceId: string    // Workspace isolation
}

Credential Types

Each credential type is defined by a credential class:

class AnthropicApi implements INodeCredential {
  label: string = 'Anthropic API'
  name: string = 'anthropicApi'
  version: number = 1.0
  
  inputs: INodeParams[] = [
    {
      label: 'Anthropic Api Key',
      name: 'anthropicApiKey',
      type: 'password'
    }
  ]
}

Credential types are loaded from flowise-components/dist/credentials at server startup, similar to how nodes are loaded.

Supported Credential Types

Flowise supports credentials for numerous services:

AI Providers

OpenAI, Anthropic, Google, Azure, Cohere, Hugging Face

Cloud Services

AWS, Azure, Google Cloud, Cloudflare

Databases

Pinecone, Supabase, MongoDB, Postgres, Astra

Search & APIs

SerpAPI, Brave Search, Serper, Tavily

Integrations

Airtable, Notion, GitHub, GitLab, Slack

Custom

MCP Servers, Custom APIs, Webhooks

Credential Lifecycle

Creating Credentials

Navigate to Credentials

Access the Credentials section in the Flowise UI

Select Type

Choose the credential type matching your service (e.g., “OpenAI API”)

Provide Details

Enter required information:

Credential name (for your reference)
API keys, tokens, or authentication details
Optional configuration parameters

Save & Encrypt

Credentials are encrypted and stored in the database

Using Credentials in Nodes

When configuring a node that requires authentication:

Node displays a “Connect Credential” dropdown
Select from your saved credentials of compatible types
Credential is linked to the node instance
At runtime, encrypted data is decrypted and injected

// In node configuration
credential: {
  label: 'Connect Credential',
  name: 'credential',
  type: 'credential',
  credentialNames: ['openAIApi', 'azureOpenAIApi'] // Compatible types
}

One credential can be reused across multiple nodes, making it easy to update API keys in one place.

Security Features

Encryption at Rest

Credentials are encrypted before storage:

Algorithm: Industry-standard encryption
Storage: Only encrypted data is persisted to the database
Decryption: Happens only at runtime when needed

Never commit the encryption key to version control. Store it securely as an environment variable.

Workspace Isolation

Credentials are scoped to workspaces:

Each credential belongs to a specific workspace
Users can only access credentials in their workspace
Prevents credential leakage between teams or projects

Flow Data Separation

Credentials are NOT included in flow exports:

// Flow data structure
{
  "nodes": [
    {
      "id": "node_123",
      "data": {
        "inputs": {
          "model": "gpt-4",
          "temperature": 0.7
        },
        "credential": "FLOWISE_CREDENTIAL_ID" // Only ID, not actual values
      }
    }
  ]
}

When importing flows, you must reconnect credentials manually.

Credential Management

Updating Credentials

To update an existing credential:

Navigate to the credential in the UI
Edit the encrypted fields
Save changes
All nodes using this credential automatically use the updated values

No need to update individual nodes when rotating API keys - just update the credential once.

Deleting Credentials

Deleting a credential that’s in use will break associated nodes. Flowise may show warnings if credentials are referenced by active chatflows.

Before deletion:

Check which chatflows use the credential
Update or remove those references
Then safely delete the credential

Credential Icons

Credentials inherit icons from their associated nodes:

// During node initialization
if (newNodeInstance.credential) {
  for (const credName of newNodeInstance.credential.credentialNames) {
    this.credentialIconPath[credName] = nodeIconAbsolutePath
  }
}

// During credential initialization
newCredInstance.icon = this.credentialIconPath[newCredInstance.name] ?? ''

This provides visual consistency in the UI.

Advanced Features

Multi-Field Credentials

Some credentials require multiple values:

inputs: [
  {
    label: 'AWS Access Key ID',
    name: 'accessKeyId',
    type: 'password'
  },
  {
    label: 'AWS Secret Access Key',
    name: 'secretAccessKey',
    type: 'password'
  },
  {
    label: 'AWS Region',
    name: 'region',
    type: 'string',
    default: 'us-east-1'
  }
]

OAuth Credentials

For services requiring OAuth flows:

Credential type includes OAuth endpoints
User initiates authentication flow
Access and refresh tokens are stored encrypted
Tokens are automatically refreshed when expired

Conditional Credentials

Some nodes accept multiple credential types:

credentialNames: ['openAIApi', 'azureOpenAIApi', 'openRouterApi']

Users choose which type to use based on their provider.

API Integration

Credential Request Flow

When creating via API:

interface ICredentialReqBody {
  name: string                    // Display name
  credentialName: string          // Type (e.g., "anthropicApi")
  plainDataObj: ICredentialDataDecrypted // Unencrypted values
  workspaceId: string
}

Server encrypts plainDataObj before storage.

Credential Response

When retrieved (for editing):

interface ICredentialReturnResponse extends ICredential {
  plainDataObj: ICredentialDataDecrypted // Decrypted for display
}

Decrypted credentials are only returned to authorized users and should never be logged or transmitted insecurely.

Best Practices

Security Best Practices

Use unique credentials for each environment (dev, staging, prod)
Rotate API keys regularly
Use least-privilege access (API keys with minimal required permissions)
Monitor credential usage through analytics
Delete unused credentials to reduce attack surface

Organization Best Practices

Use descriptive names: “Production OpenAI Key” vs “OpenAI 1”
Document credential purpose in the name or a comment system
Create separate credentials for different projects/teams
Audit credential usage periodically

Credential Types Reference

Common Credential Types

OpenAI API (openAIApi)

OpenAI API Key
Optional: Organization ID

Anthropic API (anthropicApi)

Anthropic API Key

Google Generative AI (googleGenerativeAI)

Google API Key

Pinecone (pineconeApi)

Pinecone API Key
Environment (deprecated in newer versions)

Supabase (supabaseApi)

Supabase URL
Supabase API Key

AWS (awsCredential)

Access Key ID
Secret Access Key
Region

MCP Server (mcpServer)

Server configuration (command, args, env)

Testing Credentials

While Flowise doesn’t have built-in credential testing, you can verify credentials by:

Creating a simple test chatflow
Using the credential in a basic node
Running a test query
Checking for authentication errors

Troubleshooting

Credential Not Appearing in Dropdown

Ensure credential type matches node’s credentialNames
Check that credential was saved successfully
Verify workspace permissions

Authentication Errors at Runtime

Verify API key is correct and active
Check API key permissions/scopes
Confirm service is not experiencing outages
Review rate limits and quotas

Credential Lost After Import

Credentials are not included in exports (security feature)
Manually recreate credentials in the target environment
Reconnect credentials to imported nodes

Nodes - Using credentials in node configurations
Chatflows - Building workflows with authenticated nodes
Agentflows - Using credentials in autonomous workflows

Get Started

Core Concepts

Building Flows

Deployment

Configuration

​What are Credentials?

​Credential Architecture

​Credential Structure

​Credential Types

​Supported Credential Types

AI Providers

Cloud Services

Databases

Search & APIs

Integrations

Custom

​Credential Lifecycle

​Creating Credentials

​Using Credentials in Nodes

​Security Features

​Encryption at Rest

​Workspace Isolation

​Flow Data Separation

​Credential Management

​Updating Credentials

​Deleting Credentials

​Credential Icons

​Advanced Features

​Multi-Field Credentials

​OAuth Credentials

​Conditional Credentials

​API Integration

​Credential Request Flow

​Credential Response

​Best Practices

​Credential Types Reference

​Common Credential Types

​Testing Credentials

​Troubleshooting

​Related Concepts

Build docs developers (and LLMs) love

What are Credentials?

Credential Architecture

Credential Structure

Credential Types

Supported Credential Types

Credential Lifecycle

Creating Credentials

Using Credentials in Nodes

Security Features

Encryption at Rest

Workspace Isolation

Flow Data Separation

Credential Management

Updating Credentials

Deleting Credentials

Credential Icons

Advanced Features

Multi-Field Credentials

OAuth Credentials

Conditional Credentials

API Integration

Credential Request Flow

Credential Response

Best Practices

Credential Types Reference

Common Credential Types

Testing Credentials

Troubleshooting

Related Concepts