MCP / OAuth 2.0

Connect Telara from your AI platform

Choose Telara in your platform's connector gallery, select Connect, and sign in as yourself. You do not need to install Telara, paste an API key, enter an MCP URL, or choose a tenant.

One-click connection

OAuth 2.0 + PKCE

Dynamic client registration

Employee-scoped access

Employee setup

Connect in three steps

1. Select Connect

Open your AI platform's connector gallery, choose Telara, and select Connect.

2. Sign in to Telara

Use your existing employee account. Telara resolves your organization from that identity; the AI platform never supplies a tenant ID.

3. Approve and return

Review the requested scope, approve it, and return to your AI platform. The connection is now available to that employee.

Requirements

What must already be available

An active Telara employee account in an organization that uses Telara.
Access to at least one company system or knowledge source granted through that organization's Telara policies.
Permission from the AI platform or its administrator to use third-party connectors. Some enterprise platforms require an administrator to enable a connector before employees can see it.

No shared Telara credential is created. Suspending the employee, removing their organization membership, revoking the OAuth grant, or disconnecting Telara stops future access.

Tools and permissions

The platform receives only the tools that employee may use

Search and fetch

Read-only tools search approved connected systems, retrieve permitted records, and return source context. Each tool publishes an explicit title and safety annotations, including readOnlyHint.

Actions

Platforms that support actions receive the generic telara_execute_action tool. They first read telara://integrations/available for the employee's allowed actions, then call it with an integration, action, and parameters. Telara still enforces the employee's policy, approval requirements, and source permissions before an action proceeds. Read-only connector galleries ignore write tools.

Publisher review

Reviewer setup and verification

1. Connect with OAuth

In the platform under review, add Telara using the published connector entry or the MCP endpoint below. Select Connect, sign in with the reviewer account supplied privately by Telara, and approve the requested access. No API key, tenant ID, or manually selected organization is used.

2. Confirm the advertised tool set

After OAuth completes, the client should receive exactly these platform tools for a policy-enabled reviewer account:

telara_knowledge_search, telara_traverse, telara_get_context, telara_impact, telara_annotate, telara_link, telara_timeline, telara_code_call_hierarchy, telara_browser_extract, telara_archive_read, telara_archive_ls, telara_archive_search, telara_task_list, telara_task_create, telara_task_resume, telara_task_checkpoint, telara_task_complete, telara_task_pause, telara_execute_action, telara_tool_search, telara_tool_describe

The tool list is employee and policy scoped. A reviewer with a different policy can receive a smaller set; Telara does not expose another tenant's tools or data.

3. Run the review prompts

Ask: “Search my permitted Telara knowledge for the reviewer welcome record and cite the source.”
Ask: “Find an authorized action for the reviewer test integration.” Telara should use telara_tool_search; the reviewer may inspect the returned action with telara_tool_describe.
Where the platform permits writes, execute only the supplied non-destructive reviewer test action through telara_execute_action. It remains subject to Telara policy and any configured approval requirement.
Send the approximate test time and platform name to support if verification is needed. Telara can confirm the client, tool, result, and policy decision from its audit trail; never send passwords, access tokens, or copied customer data.

Support and privacy

Help for users and platform reviewers

For connector setup, reviewer-account provisioning, security reports, or technical support, email support@telara.dev. For privacy requests, email privacy@telara.dev.

Read Telara's Privacy Policy and Terms of Service.

Protocol details

Standards-based discovery and authorization

MCP endpoint: https://api.telara.dev/v1/mcp
Authorization metadata: https://api.telara.dev/.well-known/oauth-authorization-server
Protected resource metadata: https://api.telara.dev/.well-known/oauth-protected-resource
OAuth flow: Authorization code with PKCE S256, refresh tokens, revocation, and RFC 7591 dynamic client registration.

These values are for platform publishers and troubleshooting. Employees connecting from a published gallery should not enter them manually.

Troubleshooting

If Connect does not finish

The connector is unavailable or disabled

Ask your AI platform administrator to enable Telara for your organization or group. Platform approval is separate from your Telara permissions.

Sign-in succeeds but no sources appear

Confirm that the employee belongs to the intended Telara organization and has access to at least one connected source. Telara never lets the client select or override a tenant.

Authorization loops or returns an error

Disconnect Telara in the AI platform, then connect again. If the problem continues, send the time of the attempt and the platform name to support@telara.dev. Never send a password or access token.