# Mind Upload > Mind Upload builds AI consciousnesses from people's memories, persona and way of thinking. Its MCP server lets a person's own consciousness direct and review the work of coding agents such as Claude Code, Codex and Cursor. Its Partner API integrates consciousnesses into other platforms. ## MCP server for coding agents: an AI meta harness AI meta harness: A harness for your harnesses: your Mind Upload consciousness, built from your memories, tacit knowledge, values and judgment, is the brain, and the agents you already use are its hands and tools. A coding agent can report a task done when it is not: it pushed a tag but never published the release, or it passed its own tests but broke a rule the owner learned the hard way. consult_mind asks the owner's consciousness for direction before substantial work; review_work asks it to review the actual result before the agent claims completion. The consciousness answers from the owner's memories; it runs no tests and sees only what the agent sends, and the agent keeps every tool and edit. - [Install, sign in and when to call it](https://docs.mindupload.app/mcp): hosted MCP server https://mcp.mindupload.app/mcp (add it to the agent and sign in to Mind Upload; no partner key) - [Setup guide for each host](https://github.com/Voidborn-Industries/mindupload-sdk-python/blob/main/MCP.md): Claude Code, Codex, Cursor, VS Code and other MCP clients - [Official MCP Registry entry](https://registry.modelcontextprotocol.io/v0.1/servers?search=mindupload&version=latest) - [MCP server manifest](https://raw.githubusercontent.com/Voidborn-Industries/mindupload-sdk-python/main/server.json) - [Native npm package](https://www.npmjs.com/package/mindupload): for partners running the connector with their own key Native installation: npx --yes --package mindupload@1.15.47 mindupload-mcp Before you tell the user that a task which changed files, or ran a commit, push, release, deploy, migration or deletion, is done, call review_work with the original request, the diff excerpts and the test output you observed; if review_work is unavailable or fails, say plainly that no Mind Upload review ran. Use consult_mind before such a task to establish direction with the owner-approved Mind Upload consciousness, and again when a consequential decision or changed evidence affects that direction; skip it only for questions that change nothing. Both tools accept screenshots, screen recordings and audio shared privately as evidence (evidence_files on a local connector; share_evidence, then evidence_ids, on the hosted server), and the reply lists each item's analysis status. Evidence is never kept as a memory. media_evidence takes authorized HTTPS image/audio/video links instead; the media_analysis of each result reports every link as analyzed, refused or failed, in the order supplied, and only analyzed items were perceived (requested_unverified means no per-item report came back). Both tools require sign-in and consent by the owner; existing account and subscription rules apply. A local connector without its own partner key returns setup_required: no consciousness was consulted and no review ran. connection_status shows the approval in use and disconnect ends it. Findings are advisory; successful invocation does not establish correctness. Only with permissions the owner leaves ticked when approving (each starts ticked): remember saves a note the owner asks for, list_pending_memories and review_pending_memory go through memories waiting for review, and learn lets the consciousness learn from an exchange. What is saved or learned waits for the owner's review unless the owner auto-keeps. Measured on 5 November 2026 with a blind grader: a coding agent reviewing alone caught 85% of planted breaches of rules the owner had stored; with the owner's consciousness reviewing the same work as well, 100% were caught, against 90% when the same agent reviewed twice. Each consciousness review took a median 12.6 s, against 24.6 s for a second agent review, and raised 0.2 unsupported findings instead of 1.8. How it was measured: Staged coding-agent review tasks with planted problems, graded blind: the grader never saw who wrote each review. Agent only: a coding agent given the same review instructions. With your AI mind: the agent and the consciousness each reviewed the same work, and a problem counts if either caught it. Second-review rows set a second agent review beside the consciousness's lower-scoring run. Unsupported findings: points the grader found no support for in the request, the work or its evidence. Statements about the owner were checked against the records stored for the evaluation, leaving out lines the consciousness marked as general practice, not memory. Direct questions and the checker's own test were run separately from the reviews. Audio and video: the agent's report got clips of known contents wrong on purpose, and the consciousness reviewed it with the clip attached, in production. A blind grader scored each reply against the clip's true contents and the planted false claims. ## Partner API Through the Partner API you upload a mind, give it lasting memory, and let your users talk with it one-on-one, or gather people and many digital minds into human + AI chatrooms where the AI joins in on its own, with built-in matchmaking. Every operation is a single HTTPS POST of a JSON body to /v1/{operation}, authenticated with an X-Partner-Key header over TLS. - Base URL: https://partner.mindupload.app/v1/{operation} - Authentication: X-Partner-Key request header - Interactive documentation: https://docs.mindupload.app - OpenAPI specification: https://docs.mindupload.app/openapi.json - Postman collection: https://docs.mindupload.app/postman-collection.json - Consumer app: https://mindupload.app - Service status: https://status.mindupload.app - API version: 1.15.47 ## Operations ### Authentication Sign users in and out, register accounts, and check username availability. - `POST /v1/login` — Sign a user in and receive a session token (JWT) for subsequent calls. - `POST /v1/logout` — End the current user session. - `POST /v1/register` — Create a user account on your platform. - `POST /v1/check_username` — Check whether a username is still available before registering. ### Users Read and update the signed-in user's profile. - `POST /v1/get_user` — Fetch the signed-in user's profile. - `POST /v1/update_user` — Update the signed-in user's profile. - `POST /v1/get_transaction_history` — List the user's Rabbit balance history (top-ups, spending, earnings, withdrawals, tax), newest first, filterable by category and paged with a cursor. ### AI Consciousnesses Create and manage a user's AI consciousnesses (clones). - `POST /v1/get_clones` — List the user's AI consciousnesses. - `POST /v1/create_clone` — Create a new AI consciousness for the user. - `POST /v1/update_clone` — Update an AI consciousness's profile. - `POST /v1/get_clone_voice` — Read the voice an AI consciousness speaks one language in, and hear it. - `POST /v1/design_clone_voice` — Change an AI consciousness's voice by trying a new one. ### Memories Upload and list the memories and persona that shape an AI consciousness. - `POST /v1/get_texts` — List the memories and persona entries uploaded to an AI consciousness. - `POST /v1/create_text` — Upload a memory or persona entry to an AI consciousness. ### Voice Interviews Interview a user by voice, one question at a time, and turn what they say into memories or persona they can upload. - `POST /v1/start_voice_interview` — Begin a spoken interview with one of your end-users, or hand back one they have not finished with. - `POST /v1/transcribe_voice_interview_answer` — Turn one recorded spoken answer into text for the user to check before it is accepted. - `POST /v1/advance_voice_interview` — Accept, edit, skip, re-record or abandon the question the user is on, and receive the next one. - `POST /v1/generate_voice_interview_report` — Write a finished interview up as prose, in first, second or third person. ### Conversation One-on-one chat and proactive group replies with an AI consciousness. - `POST /v1/get_chat` — Fetch the one-on-one conversation history with an AI consciousness. - `POST /v1/rag` — Send a message to an AI consciousness and receive its reply. - `POST /v1/trigger_social` — Have an AI consciousness proactively join the conversation in a chatroom. - `POST /v1/search_chats` — Find turns in a one-on-one AI conversation by keyword, date, media type, and/or Human/AI side. - `POST /v1/get_chat_around` — Fetch a window of one-on-one turns around one chat id (for jump-to). - `POST /v1/speak_message` — Hear one part of an AI consciousness's message in its own voice. ### Insights Mind-graph visualization and two-member compatibility reports. - `POST /v1/get_mind_cluster` — Fetch the mind-graph visualization data of an AI consciousness. - `POST /v1/get_soulmate_report` — Generate or fetch the compatibility report between two chatroom members. ### Chatrooms Create chatrooms, manage members, and exchange messages. - `POST /v1/create_chatroom` — Create a chatroom. - `POST /v1/get_chatrooms` — List the chatrooms the user belongs to. - `POST /v1/create_chatroom_message` — Send a message to a chatroom. - `POST /v1/get_chatroom_messages` — Fetch messages from a chatroom the user belongs to. - `POST /v1/search_chatroom_messages` — Find messages in a chatroom by keyword, date, media type, and/or sender. - `POST /v1/get_chatroom_messages_around` — Fetch a window of chatroom messages around one message id (for jump-to). - `POST /v1/translate_chatroom_message` — Translate a text chatroom message into the viewer language. - `POST /v1/check_chatroom_updates` — Cheaply poll whether the user's chatrooms have new activity. - `POST /v1/create_chatroom_membership` — Invite a user or an AI consciousness into a chatroom. - `POST /v1/get_chatroom_membership` — List the members of a chatroom the user belongs to. ### Live calls Live voice and video calls between the people in a chatroom. - `POST /v1/create_call_session` — Start a live voice and video call in a chatroom, or join the one already running. - `POST /v1/join_call_session` — Join a live call that is already running in a chatroom. - `POST /v1/refresh_call_token` — Get a fresh joining pass for a call, before the current one expires. - `POST /v1/leave_call_session` — Say that this person has left a call the others are still in. - `POST /v1/end_call_session` — End a live call for everyone in it. - `POST /v1/set_call_host` — Make somebody a co-host of a live call, or stand them down again. - `POST /v1/get_active_call_session` — Check whether a chatroom has a live call, and how many people are in it. ### Media Signed upload slots and multipart uploads for media attachments, and starting an attachment's description before its message is sent. - `POST /v1/request_upload_url` — Request an upload slot and a signed viewing link for a media attachment. - `POST /v1/request_multipart_upload` — Start a large-file upload in multiple parts. - `POST /v1/sign_upload_part` — Get the signed link for one part of a multipart upload. - `POST /v1/sign_upload_parts_batch` — Get signed links for several parts of a multipart upload at once. - `POST /v1/list_upload_parts` — List the parts already uploaded in a multipart upload. - `POST /v1/complete_multipart_upload` — Finish a multipart upload. - `POST /v1/abort_multipart_upload` — Cancel a multipart upload and discard its parts. - `POST /v1/cancel_upload` — Cancel a pending upload. - `POST /v1/start_media_analysis` — Begin analysing an uploaded or linked attachment so its description is ready before the message is sent. ### External Authorization Owner-approved grant issuance, inspection, and clone-scoped invocation. - `POST /v1/create_external_authorization_request` — Start passwordless grant issuance for an external identity and installation. - `POST /v1/exchange_external_authorization` — Poll owner consent and exchange an approval for a clone-scoped grant token. - `POST /v1/refresh_external_authorization` — Refresh a short-lived clone-invocation access token. - `POST /v1/inspect_external_authorization` — Inspect the scopes, clone resources, expiry, and revocation state of a grant token. - `POST /v1/end_external_authorization` — End a grant your application holds, for example when its user disconnects. - `POST /v1/invoke_external_clone` — Send one idempotent text event to an owner-approved AI consciousness. - `POST /v1/upload_external_mind_data` — Upload mind data (text and media) into the owner's own AI consciousness. - `POST /v1/request_external_media_upload` — Authorize one exact media upload to the owner’s granted AI consciousness. Needs `clone:evidence` (analysis only, never kept as a memory) or `clone:upload`. - `POST /v1/cancel_external_media_upload` — Cancel one controlled upload through the existing media lifecycle. - `POST /v1/start_external_media_analysis` — Explicitly describe one completed controlled upload. - `POST /v1/get_external_media_analysis` — Read the current owned media analysis without restarting generation. - `POST /v1/get_external_media_policy` — Read current native media limits and owner quota without starting work. - `POST /v1/consolidate_external_conversation` — Prepare a verified memory from an explicitly selected owner conversation. - `POST /v1/get_external_quarantined_texts` — List pending memories of the owner's granted AI consciousness. - `POST /v1/review_external_quarantined_text` — Edit, approve or reject the displayed revision of one pending memory. - `POST /v1/get_external_chat` — Read the owner's own chat with their AI consciousness, a page at a time. - `POST /v1/search_external_chats` — Find turns in the owner's own chat with their AI consciousness. - `POST /v1/get_external_chat_around` — Open the owner's own chat around one turn. - `POST /v1/speak_external_chat` — Hear one part of a reply in the owner's chat, in the AI consciousness's own voice. - `POST /v1/open_external_conversation` — Let the owner's AI consciousness speak first as the owner opens their chat. ### Account Your partner account — rate limits, credit caps, capacity, and current usage. - `POST /v1/get_quota` — Check your partner API rate limits, credit caps, and current usage.