Integrations

Use Saywise from your AI tools

Saywise runs an MCP server, so the agent you already work in can post your AI work and usage stats straight to your profile. Everything lands as a private draft you review before anything goes public.

Claude Code

Add the server once, then authenticate from inside a session:

claude mcp add saywise --transport http --url https://saywise.com/api/mcp/v1

Run /mcp in Claude Code and pick saywise — your browser opens to sign in and approve the connection.

Claude Desktop and claude.ai

Open Settings → Connectors → Add custom connector.
Paste the server URL:

https://saywise.com/api/mcp/v1

Other agents

Any client that speaks MCP over Streamable HTTP can connect — point it at the URL above. Auth is standard OAuth 2.1 with dynamic client registration, so there are no API keys to manage; the client opens a browser for consent on first use.

What the tools do

saywise_create_suggested_drafts— your agent sends the session’s content and Saywise generates story drafts for your profile.
saywise_get_story_scan_instructions — guides your agent through mining past conversations for work worth telling.
saywise_get_usage_stats_instructions and saywise_submit_usage_stats — measure your AI usage locally and, once you approve the exact numbers, add them to your AI Stack.

Start with a prompt

Connected? Paste this into your agent to get your first drafts:

I want to turn my past AI sessions into Saywise story drafts. Go through my recent sessions, shortlist the ones where something real shipped, got fixed, or got measured, and let me pick. Then send the picks to Saywise with its tools — it generates the drafts — and give me the review link. Skip anything sensitive, and keep everything as drafts — I'll approve them myself.

Privacy

Conversation scans run on your machine or in your account — only the content of sessions you pick is sent to Saywise, so it can generate the drafts.
Drafts stay private until you approve them yourself.
Usage stats are aggregate numbers only — counts, dates, model names — and are submitted only after your agent shows you the exact payload and you say yes.

What the tools do

saywise_create_suggested_drafts— your agent sends the session’s content and Saywise generates story drafts for your profile.

saywise_get_story_scan_instructions — guides your agent through mining past conversations for work worth telling.

saywise_get_usage_stats_instructions and saywise_submit_usage_stats — measure your AI usage locally and, once you approve the exact numbers, add them to your AI Stack.

Start with a prompt

Connected? Paste this into your agent to get your first drafts:

I want to turn my past AI sessions into Saywise story drafts. Go through my recent sessions, shortlist the ones where something real shipped, got fixed, or got measured, and let me pick. Then send the picks to Saywise with its tools — it generates the drafts — and give me the review link. Skip anything sensitive, and keep everything as drafts — I'll approve them myself.

Privacy

Conversation scans run on your machine or in your account — only the content of sessions you pick is sent to Saywise, so it can generate the drafts.

Drafts stay private until you approve them yourself.

Usage stats are aggregate numbers only — counts, dates, model names — and are submitted only after your agent shows you the exact payload and you say yes.