ChatGPT MCP server for AI agents
Let Claude, ChatGPT or your own agent read and act on ChatGPT: 33 typed functions through one MCP server. No ChatGPT API key needed.
- Site
- chatgpt.com
- Functions
- 33
- Sign-in
- For account actions
- Works with
- MCP, HTTP, Node, Python
Connect
An MCP server for ChatGPT alone, with one tool per function. Paste it into Claude, ChatGPT, Cursor or any MCP client as a remote server.
https://api.bowmark.ai/mcp/site/chatgptOne MCP server for ChatGPT and every other site Bowmark covers. Your agent looks up what it needs and calls it.
https://api.bowmark.ai/mcpSet it up in your clientBoth need a Bowmark API key. The quickstart gets you one and a first result in a few minutes. Functions that need a login use the person's own account: how sign-in works.
Functions
Open a function to see its arguments, what it returns and how to call it.
browseGPTStoreLists the GPT Store front page by category.Arguments
None.
Returns
GPTStoreSection[]interface GPTSummary { id: string; name: string | null; description: string | null; author: string | null; url: string | null; /** The site's own rounded count string, e.g. "1M+". */ conversations: string | null; } interface GPTStoreSection { id: string | null; title: string | null; gpts: GPTSummary[]; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/browseGPTStoreCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.browseGPTStore();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.browseGPTStore())curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/browseGPTStore \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.browseGPTStore();getAccountReads the CALLER's account.Arguments
None.
Returns
ChatGPTAccountinterface ChatGPTAccount { id: string | null; email: string | null; name: string | null; plan: string | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/getAccountCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.getAccount();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.getAccount())curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/getAccount \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.getAccount();getConversationReads one of the CALLER's conversations in full.Arguments
Name Type Required idOrUrlstringYes Returns
Conversationinterface ConversationMessage { id: string; role: string; text: string; createTime: string | null; /** The model that answered, on assistant turns. */ model: string | null; } interface Conversation { id: string; title: string | null; createTime: string | null; updateTime: string | null; gptId: string | null; /** The visible thread, oldest first. */ messages: ConversationMessage[]; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/getConversationCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.getConversation("6956c161-a7a4-8001-bc75-894ecaaa9ae6");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.getConversation("6956c161-a7a4-8001-bc75-894ecaaa9ae6"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/getConversation \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"idOrUrl":"6956c161-a7a4-8001-bc75-894ecaaa9ae6"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.getConversation("6956c161-a7a4-8001-bc75-894ecaaa9ae6");getCustomInstructionsReads the CALLER's personalization.Arguments
None.
Returns
CustomInstructionsinterface CustomInstructions { enabled: boolean; nickname: string | null; occupation: string | null; traits: string | null; aboutYou: string | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/getCustomInstructionsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.getCustomInstructions();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.getCustomInstructions())curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/getCustomInstructions \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.getCustomInstructions();getGPTReads one custom GPT's page in the GPT Store.Arguments
Name Type Required idOrUrlstringYes Returns
GPTDetail | nullinterface GPTSummary { id: string; name: string | null; description: string | null; author: string | null; url: string | null; /** The site's own rounded count string, e.g. "1M+". */ conversations: string | null; } interface GPTDetail extends GPTSummary { conversationStarters: string[]; categories: string[]; rating: number | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/getGPTCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.getGPT("g-2fkFE8rbu");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.getGPT("g-2fkFE8rbu"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/getGPT \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"idOrUrl":"g-2fkFE8rbu"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.getGPT("g-2fkFE8rbu");getMemoriesReads what ChatGPT has saved to the CALLER's memory (Settings → Personalization → Manage memories).Arguments
None.
Returns
Memoriesinterface Memories { memories: { id: string; content: string; updateTime: string | null }[]; /** Share of the memory budget used, 0-100, when reported. */ percentFull: number | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/getMemoriesCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.getMemories();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.getMemories())curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/getMemories \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.getMemories();getProjectReads one of the CALLER's projects.Arguments
Name Type Required idOrUrlstringYes opts{ cursor?: string }No Returns
ProjectDetailinterface ProjectConversation { id: string; title: string | null; updateTime: string | null; } interface ProjectDetail { id: string; name: string | null; instructions: string | null; files: { id: string; name: string | null }[]; conversations: ProjectConversation[]; /** Pass back as cursor for more; null at the end. */ cursor: string | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/getProjectCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.getProject("g-p-6956c161a7a48001");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.getProject("g-p-6956c161a7a48001"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/getProject \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"idOrUrl":"g-p-6956c161a7a48001"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.getProject("g-p-6956c161a7a48001");listConversationsLists the CALLER's ChatGPT chat history, newest first.Arguments
Name Type Required opts{ limit?: number; offset?: number; archived?: boolean }No Returns
ConversationListinterface ConversationSummary { id: string; title: string | null; createTime: string | null; updateTime: string | null; gptId: string | null; projectId: string | null; } interface ConversationList { conversations: ConversationSummary[]; total: number | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/listConversationsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.listConversations({});import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.listConversations({}))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/listConversations \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"opts":{}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.listConversations({});listLibraryImagesLists the images the CALLER has generated, from the Library page, newest first, with the conversation each came from.Arguments
None.
Returns
LibraryImage[]interface LibraryImage { id: string; url: string | null; prompt: string | null; conversationId: string | null; createTime: string | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/listLibraryImagesCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.listLibraryImages();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.listLibraryImages())curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/listLibraryImages \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.listLibraryImages();listModelsLists the models the CALLER's plan can pick in the model menu, with their limits.Arguments
None.
Returns
ChatGPTModel[]interface ChatGPTModel { /** The id the model menu sends, e.g. "gpt-4o". */ slug: string; title: string | null; description: string | null; maxTokens: number | null; tags: string[]; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/listModelsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.listModels();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.listModels())curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/listModels \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.listModels();listMyGPTsLists the custom GPTs in the CALLER's sidebar (their own and ones they pinned or used), with visibility and pinned state.Arguments
None.
Returns
MyGPT[]interface GPTSummary { id: string; name: string | null; description: string | null; author: string | null; url: string | null; /** The site's own rounded count string, e.g. "1M+". */ conversations: string | null; } interface MyGPT extends GPTSummary { /** "only me", "link" or "store" when the site reports it. */ visibility: string | null; pinned: boolean; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/listMyGPTsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.listMyGPTs();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.listMyGPTs())curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/listMyGPTs \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.listMyGPTs();listProjectsLists the CALLER's ChatGPT projects with their names and conversation counts.Arguments
None.
Returns
ProjectSummary[]interface ProjectConversation { id: string; title: string | null; updateTime: string | null; } interface ProjectSummary { /** g-p-… id; pass to getProject. */ id: string; name: string | null; conversations: ProjectConversation[]; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/listProjectsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.listProjects();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.listProjects())curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/listProjects \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.listProjects();listScheduledTasksLists the CALLER's scheduled tasks (reminders and recurring prompts ChatGPT runs on a schedule) with their next run time.Arguments
None.
Returns
ScheduledTask[]interface ScheduledTask { id: string; title: string | null; prompt: string | null; /** The recurrence the site stores, e.g. "FREQ=DAILY". */ schedule: string | null; nextRun: string | null; enabled: boolean; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/listScheduledTasksCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.listScheduledTasks();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.listScheduledTasks())curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/listScheduledTasks \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.listScheduledTasks();searchConversationsSearches the CALLER's chat history by keyword, the way the sidebar's search does, and returns matching conversations with a snippet.Arguments
Name Type Required querystringYes opts{ cursor?: string }No Returns
ConversationSearchResultinterface ConversationSearchHit { id: string; title: string | null; updateTime: string | null; /** The matching text, when the site returns one. */ snippet: string | null; } interface ConversationSearchResult { hits: ConversationSearchHit[]; /** Pass back as `cursor` for the next page; null at the end. */ cursor: string | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/searchConversationsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.searchConversations("trip");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.searchConversations("trip"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/searchConversations \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"query":"trip"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.searchConversations("trip");searchGPTsSearches the GPT Store by keyword and returns matching custom GPTs with their id, name, description, author and conversation count.Arguments
Name Type Required querystringYes Returns
GPTSummary[]interface GPTSummary { id: string; name: string | null; description: string | null; author: string | null; url: string | null; /** The site's own rounded count string, e.g. "1M+". */ conversations: string | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/searchGPTsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.searchGPTs("writing");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.searchGPTs("writing"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/searchGPTs \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"query":"writing"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.searchGPTs("writing");
archiveConversationArchives (or un-archives) one of the CALLER's conversations.Arguments
Name Type Required conversationstringYes opts{ archived?: boolean }No Returns
WriteResultinterface WriteResult { ok: boolean; /** The conversation, memory or share id the write acted on. */ id: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/archiveConversationCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.archiveConversation("<conversation>");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.archiveConversation("<conversation>"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/archiveConversation \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"conversation":"<conversation>"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.archiveConversation("<conversation>");createGPTCreates a custom GPT in the CALLER's account.Arguments
Name Type Required draft{ name: string; description?: string; instructions?: string; promptStarters?: string[] }Yes Returns
CreatedGPTinterface CreatedGPT { /** The g-… id getGPT, pinGPT and sendMessage take. */ id: string; name: string; url: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/createGPTCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.createGPT({});import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.createGPT({}))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/createGPT \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"draft":{}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.createGPT({});createProjectCreates a new project in the CALLER's account with a name and optional instructions.Arguments
Name Type Required namestringYes Returns
CreatedProjectinterface CreatedProject { /** The g-p-… id getProject and listProjects use. */ id: string; name: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/createProjectCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.createProject("<name>");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.createProject("<name>"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/createProject \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"name":"<name>"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.createProject("<name>");deleteConversationDeletes one of the CALLER's conversations.Arguments
Name Type Required conversationstringYes Returns
WriteResultinterface WriteResult { ok: boolean; /** The conversation, memory or share id the write acted on. */ id: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/deleteConversationCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.deleteConversation("<conversation>");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.deleteConversation("<conversation>"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/deleteConversation \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"conversation":"<conversation>"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.deleteConversation("<conversation>");deleteMemoryDeletes one saved memory from the CALLER's ChatGPT memory.Arguments
Name Type Required memoryIdstringYes Returns
WriteResultinterface WriteResult { ok: boolean; /** The conversation, memory or share id the write acted on. */ id: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/deleteMemoryCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.deleteMemory("<memoryId>");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.deleteMemory("<memoryId>"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/deleteMemory \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"memoryId":"<memoryId>"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.deleteMemory("<memoryId>");generateImageAsks ChatGPT to create an image from a prompt as the CALLER and returns the image url(s) and the conversation they landed in.Arguments
Name Type Required promptstringYes Returns
GeneratedImagesinterface GeneratedImages { conversationId: string | null; /** One entry per picture; url is a short-lived download link. */ images: { fileId: string; url: string | null }[]; reply: string; warnings: string[]; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/generateImageCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.generateImage("<prompt>");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.generateImage("<prompt>"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/generateImage \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"prompt":"<prompt>"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.generateImage("<prompt>");moveConversationToProjectFiles one of the CALLER's conversations under one of their projects.Arguments
Name Type Required conversationstringYes projectstringYes Returns
WriteResultinterface WriteResult { ok: boolean; /** The conversation, memory or share id the write acted on. */ id: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/moveConversationToProjectCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.moveConversationToProject("<conversation>", "<project>");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.moveConversationToProject("<conversation>", "<project>"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/moveConversationToProject \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"conversation":"<conversation>","project":"<project>"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.moveConversationToProject("<conversation>", "<project>");pinGPTPins a GPT to (or unpins it from) the CALLER's sidebar.Arguments
Name Type Required gptstringYes Returns
WriteResultinterface WriteResult { ok: boolean; /** The conversation, memory or share id the write acted on. */ id: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/pinGPTCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.pinGPT("<gpt>");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.pinGPT("<gpt>"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/pinGPT \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"gpt":"<gpt>"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.pinGPT("<gpt>");rateResponseGives a thumbs up or thumbs down on one of ChatGPT's replies in the CALLER's conversation, the way the buttons under a reply do.Arguments
Name Type Required conversationstringYes messageIdstringYes rating"thumbsUp" | "thumbsDown"Yes Returns
WriteResultinterface WriteResult { ok: boolean; /** The conversation, memory or share id the write acted on. */ id: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/rateResponseCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.rateResponse("<conversation>", "<messageId>", "thumbsUp");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.rateResponse("<conversation>", "<messageId>", "thumbsUp"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/rateResponse \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"conversation":"<conversation>","messageId":"<messageId>","rating":"thumbsUp"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.rateResponse("<conversation>", "<messageId>", "thumbsUp");renameConversationRenames one of the CALLER's conversations.Arguments
Name Type Required conversationstringYes titlestringYes Returns
WriteResultinterface WriteResult { ok: boolean; /** The conversation, memory or share id the write acted on. */ id: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/renameConversationCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.renameConversation("<conversation>", "<title>");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.renameConversation("<conversation>", "<title>"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/renameConversation \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"conversation":"<conversation>","title":"<title>"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.renameConversation("<conversation>", "<title>");requestDataExportRequests an export of the CALLER's ChatGPT data (Settings → Data controls → Export); OpenAI emails the download link to the caller's own address.Arguments
None.
Returns
WriteResultinterface WriteResult { ok: boolean; /** The conversation, memory or share id the write acted on. */ id: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/requestDataExportCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.requestDataExport();import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.requestDataExport())curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/requestDataExport \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.requestDataExport();sendMessageSends a message as the CALLER and returns ChatGPT's reply.Arguments
Name Type Required textstringYes opts{ conversation?: string; parentMessageId?: string; model?: string; gpt?: string; temporary?: boolean }No Returns
SentMessageinterface SentMessage { conversationId: string | null; /** The assistant's message id, pass it back as parentMessageId or to rateResponse. */ messageId: string | null; reply: string; model: string | null; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/sendMessageCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.sendMessage("<text>");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.sendMessage("<text>"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/sendMessage \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"text":"<text>"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.sendMessage("<text>");updateCustomInstructionsUpdates the CALLER's personalization.Arguments
Name Type Required update{ aboutUser?: string; responseStyle?: string; enabled?: boolean }Yes Returns
WriteResultinterface WriteResult { ok: boolean; /** The conversation, memory or share id the write acted on. */ id: string; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/updateCustomInstructionsCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.updateCustomInstructions({});import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.updateCustomInstructions({}))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/updateCustomInstructions \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"update":{}}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.updateCustomInstructions({});uploadFileUploads a text file to the CALLER's own ChatGPT file library and returns its file id.Arguments
Name Type Required namestringYes contentstringYes opts{ contentType?: string }No Returns
UploadedFileinterface UploadedFile { /** The file-… id of the upload. */ fileId: string; name: string; size: number; }HTTP endpoint
POST https://api.bowmark.ai/v1/sites/chatgpt/uploadFileCall it
import { bowmark } from "@bowmark/web"; const result = await bowmark.providers.chatgpt.uploadFile("<name>", "<content>");import asyncio from bowmark_web import bowmark result = asyncio.run(bowmark.providers.chatgpt.uploadFile("<name>", "<content>"))curl -X POST https://api.bowmark.ai/v1/sites/chatgpt/uploadFile \ -H "Authorization: Bearer $BOWMARK_API_KEY" \ -H "Content-Type: application/json" \ -d '{"name":"<name>","content":"<content>"}'// The body of a run: the MCP run tool, or POST /v1/run return await bowmark.providers.chatgpt.uploadFile("<name>", "<content>");
Use it from code
- Node
npm i @bowmark/web - Python
pip install bowmark-web - HTTP
POST /v1/sites/chatgpt/<function> - OpenAPI
/v1/sites/chatgpt/openapi.json