View on GitHub
Menu

Tako MCP for Okta

View repository

Give your AI toolsOkta context.

Connect MCP-compatible clients to your Okta environment. Discover operations, retrieve identity data, and export results through a server you deploy.

View on GitHub Explore the architecture
Your AI client

Ask.
Investigate.
Build.AI client

Claude Desktop · Cursor
VS Code · Compatible agents

STDIO / HTTP

Tako MCP

Discover → Retrieve → Export

Customer-deployed server
Okta API
Your Okta tenant

Identity
in context.Okta

Users · Groups · Applications
System logs · Policies

One connection between your AI workflow and your Okta data.

Discover what you need.
Retrieve what matters.

Start with operation details and sample data. Generate code for the retrieval, then return useful summaries and CSV exports rather than putting every raw record into the conversation.

  1. Discover

    Understand the operation.

    Find available operations and inspect endpoint details before retrieving the full results.

    get_available_operations
  2. Retrieve

    Run the retrieval.

    Generate Python for the query. The server validates code before executing the configured workflow.

    execute_code
  3. Export

    Keep the result useful.

    Return summaries to the client and save larger result sets as CSV for reporting and further analysis.

    CSV + summary

The diagram describes Agent mode. Result size and speed depend on Okta limits, permissions, configuration, and the query.

Choose how your tools connect.

Two operating modes. Two transport choices. Use the configuration that fits your client and deployment.

Read the configuration guide
Basic MCP mode
Expose the standard tools directly for granular tool composition.
Agent mode
Discover operations progressively for queries that need several steps.
STDIO
Connect local desktop clients through standard input and output.
HTTP
Connect remote clients through the configured MCP endpoint. Review access controls before exposing it.

Work across your identity environment.

Investigate the relationships between users, access, applications, and events with the operations exposed by the server.

Users and factors

Find user profiles, group memberships, assigned applications, and enrolled authentication factors.

Groups and applications

Inspect groups, members, application assignments, and application details.

Logs, policies, and zones

Query system log events and inspect supported policy rules and network zones.

Deploy deliberately.
Understand the boundaries.

Run the server with Docker or Python. Configure Okta credentials, permissions, client access, and AI data handling for your organization.

Okta authentication

Use an API token or OAuth 2.0 with Private Key JWT. Access depends on the configured permissions and scopes.

Code validation

AST checks restrict generated code, imports, functions, and HTTP operations. Review repository controls and deployment isolation; code validation alone is not a blanket security guarantee.

AI data processing

Okta data is sent to the AI model during queries. Larger results can be exported to CSV with summaries returned to the model. Review provider terms and local result retention.

Tako MCP is a separately licensed, customer-deployed project. Its repository documentation and operator configuration govern the deployment. Fctr’s hosted Identity Platform assurances do not apply unless an executed agreement expressly includes it.

See the connection in action.

A recorded Claude Desktop walkthrough of the server’s query workflow.

Existing demonstration recording. Interface and setup details may differ from the current repository.

Bring Okta into your AI workflow.

Start with the deployment and client setup guides.

Get Tako MCP on GitHub
Deployment guide Client setup Prefer a ready-made query interface? Explore Tako AI