Everything TalkSense does, step by step.

Four things get you from signup to your first AI-drafted post: connect an account, build a brand kit, let generation run, and review what lands in your queue.

talksense.idey.click/app/dashboard

Compose supports three modes — a single post, a thread, or an A/B test with two variants — and six content shapes: standard, list, news reaction, poll, quote, and thread. Generate with AI, write manually, or start from a saved draft template.

talksense.idey.click/app/compose

Everything — AI-generated, manually written, or handed off by an MCP client — lands in the Approval Queue as pending, unless auto-execute is on for that action type and the AI's confidence clears your threshold. From the queue you can:

• Edit content inline before approving
• Approve or reject individually, or in bulk
• See live engagement metrics once a post is published
• Filter by account, type, or status

TalkSense runs an MCP server — the open protocol Claude, ChatGPT, and other AI assistants use to call external tools. Once connected, your assistant can look up your accounts, read your brand kit, and hand a finished draft straight to your TalkSense approval queue — using the subscription you already pay for, instead of TalkSense's own generation credits.

1

Generate an API key

In TalkSense, go to Settings → Integrations. Under API Keys, name the client you're connecting — e.g. Claude Desktop — and click Generate Key.

The key is shown exactly once. Copy it now — TalkSense never displays it again, only a prefix to help you tell keys apart later.

talksense.idey.click/app/settings

2

Note your server URL

Just above the API Keys panel, MCP Connection shows your server URL — it's your TalkSense domain plus /api/v1/mcp. You'll need this and the key from step 1.

talksense.idey.click/app/settings

3

Add it to Claude Desktop or claude.ai

Current versions of Claude support remote MCP connectors directly: Settings → Connectors → Add custom connector, then paste your server URL and set the access token to the key from step 1.

If your version only reads a local config file, bridge it with mcp-remote instead:

{
  "mcpServers": {
    "talksense": {
      "command": "npx",
      "args": [
        "-y", "mcp-remote",
        "https://api.talksense.idey.click/api/v1/mcp",
        "--header", "Authorization:Bearer ${TALKSENSE_API_KEY}"
      ],
      "env": { "TALKSENSE_API_KEY": "ts_mcp_..." }
    }
  }
}

4

Add it to ChatGPT

In ChatGPT, open Settings → Connectors (menu names shift between plans and versions — look for "custom connector" or "developer mode"), create a new connector, and supply the same server URL and bearer token.

5

Try it

Ask your assistant something like:

“Check my TalkSense brand kit, then draft a post about [topic] for my main account.”

It'll call list_accounts, then get_brand_kit, then create_draft — in that order, without you naming the tools yourself.

6

Review it in your queue

Open the Approval Queue — the draft appears with a 🔌 via MCP badge and a pending status. Edit it if you want, then approve like any other draft.

Keys are revocable any time from Settings → Integrations — revoking one only breaks that client, not the others. You can have up to 10 active keys per account.