
Take the Typeza personality test in the chat and see how two socionics types get along.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Send the person's answers for one batch of a Typeza test begun with start_test, using its draft_id and key. `answers` maps each question code to 1 (statement_1 fits better) or 2 (statement_2 fits better). Returns status next_questions with the next batch or, once every question is answered, the result: status type_found with the Typeza type (name, slug, page, short description) and keep_url, or status choose_type with candidate types when the answers fit more than one equally. Answers are saved with every call, so a test can continue later: send empty answers to get the current batch back. Give the person keep_url: it opens the full result on typeza.ai and lets them save it to an account. For a signed-in person, a first result becomes their Typeza type; a retake keeps the type they have.
The key start_test returned, passed back unchanged.
Question code to 1 (statement_1 fits better) or 2 (statement_2 fits better). Send {} to get the current batch back.
The draft_id start_test returned.
This tool isn't marked read-only, so it may change something on the server or in an account you connected.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http typeza https://api.typeza.ai/mcp
Send the person's answers for one batch of a Typeza test begun with start_test, using its draft_id and key. `answers` maps each question code to 1 (statement_1 fits better) or 2 (statement_2 fits better). Returns status next_questions with the next batch or, once every question is answered, the result: status type_found with the Typeza type (name, slug, page, short description) and keep_url, or status choose_type with candidate types when the answers fit more than one equally. Answers are saved with every call, so a test can continue later: send empty answers to get the current batch back. Give the person keep_url: it opens the full result on typeza.ai and lets them save it to an account. For a signed-in person, a first result becomes their Typeza type; a retake keeps the type they have.
| Argument | Type | Description |
|---|---|---|
| key* | string | The key start_test returned, passed back unchanged. |
| answers* | object | Question code to 1 (statement_1 fits better) or 2 (statement_2 fits better). Send {} to get the current batch back. |
| draft_id* | string | The draft_id start_test returned. |
{
"type": "object",
"title": "answer_testArguments",
"required": [
"draft_id",
"key",
"answers"
],
"properties": {
"key": {
"type": "string",
"title": "Key",
"description": "The key start_test returned, passed back unchanged."
},
"answers": {
"type": "object",
"title": "Answers",
"description": "Question code to 1 (statement_1 fits better) or 2 (statement_2 fits better). Send {} to get the current batch back.",
"additionalProperties": {
"type": "integer"
}
},
"draft_id": {
"type": "string",
"title": "Draft Id",
"description": "The draft_id start_test returned."
}
}
}Record the type the person picks for their finished Typeza test: after answer_test returns status choose_type, the candidate that describes them best, or any type they say fits them better than their result. `type` accepts a Typeza name, slug or code. Returns the result as answer_test does, with keep_url, using the Typeza type name. For a signed-in person this replaces their current Typeza type.
| Argument | Type | Description |
|---|---|---|
| key* | string | The key start_test returned, passed back unchanged. |
| type* | string | A Typeza type name (Analyst), its slug or its code, any case. |
| draft_id* | string | The draft_id start_test returned. |
{
"type": "object",
"title": "choose_test_typeArguments",
"required": [
"draft_id",
"key",
"type"
],
"properties": {
"key": {
"type": "string",
"title": "Key",
"description": "The key start_test returned, passed back unchanged."
},
"type": {
"type": "string",
"title": "Type",
"description": "A Typeza type name (Analyst), its slug or its code, any case."
},
"draft_id": {
"type": "string",
"title": "Draft Id",
"description": "The draft_id start_test returned."
}
}
}Get the signed-in person's personal Typeza invite link (https://typeza.ai/i/...), a share text with the link in it, and invitee_gets, one sentence on what the invited person gets. Call it when the person wants to invite someone, for example a partner, friend or colleague, to see how they relate. The person sends the link themselves: Typeza sends no email. The link is the same every time and works for any number of people. Needs their Typeza account connected with the relationships permission.
No arguments.
{
"type": "object",
"title": "get_my_invite_linkArguments",
"properties": {}
}Get the signed-in person's own Typeza personality type: name, subtitle, quadra and its typeza.ai page. Call it when the person asks what their type is, or before explaining how they relate to someone. Needs their Typeza account connected with the profile permission. Returns status type_found with the type, or status no_type_yet when they have not found their type yet: then offer the test here with start_test, which saves the result to their account, or give them test_url. Outputs use the Typeza type name, never a code.
No arguments.
{
"type": "object",
"title": "get_my_typeArguments",
"properties": {}
}Describe one personality type: name, subtitle, quadra, key traits, a short profile in markdown and its typeza.ai page. Call it when the person asks what a type is like, or after a test result for more detail. `type` accepts a Typeza name (Analyst), slug or code in any case; the output uses the Typeza name only.
| Argument | Type | Description |
|---|---|---|
| type* | string | A Typeza type name (Analyst), its slug or its code, any case. |
{
"type": "object",
"title": "get_personality_typeArguments",
"required": [
"type"
],
"properties": {
"type": {
"type": "string",
"title": "Type",
"description": "A Typeza type name (Analyst), its slug or its code, any case."
}
}
}Explain the relationship between two personality types: the relation's name, tagline and category, a sentence saying which side each type is on, the opening of Typeza's article on it, the relation page (relationship_url), the full article for this pair (pair_url) and a disclaimer. Call it when the person asks how two types get along, for example themselves and a partner, friend or colleague. Accepts Typeza names, slugs or codes; outputs use Typeza names, and the relation is read from type_a's side. Duality is the relation for partners and friends; when the question is about work, talk about business relations. This describes typical patterns between types, not advice or a prediction for a real relationship: always pass on the disclaimer.
| Argument | Type | Description |
|---|---|---|
| type_a* | string | A Typeza type name (Analyst), its slug or its code, any case. |
| type_b* | string | A Typeza type name (Analyst), its slug or its code, any case. |
{
"type": "object",
"title": "get_relationship_betweenArguments",
"required": [
"type_a",
"type_b"
],
"properties": {
"type_a": {
"type": "string",
"title": "Type A",
"description": "A Typeza type name (Analyst), its slug or its code, any case."
},
"type_b": {
"type": "string",
"title": "Type B",
"description": "A Typeza type name (Analyst), its slug or its code, any case."
}
}
}List the people in the signed-in person's Typeza circle, newest first. For each: the name the person knows them by (null when none is set), their label (partner, friend, family, colleague or other) when set, their Typeza type when they have one, the relationship type between the two (name, tagline, category, a sentence saying who is on which side, the relation page, the pair article on typeza.ai and whether the person's full write-up of it is open) when both have a type, and person_url, the page on typeza.ai for that relationship, where the full write-up can also be opened. Call it when the person asks about the people in their life on Typeza or how they get along with someone in their circle. Needs their Typeza account connected with the relationships permission. It never contains anyone's email, answers or interview text. Duality is the relation for partners and friends; when the question is about work, talk about business relations. This describes typical patterns between types, not advice or a prediction for a real relationship: always pass on the disclaimer. When the circle is empty, offer get_my_invite_link.
No arguments.
{
"type": "object",
"title": "list_my_peopleArguments",
"properties": {}
}List the 16 Typeza personality types: name, subtitle, quadra and the typeza.ai page of each. Call it when the person wants an overview of the types or you need the exact Typeza names. Outputs use Typeza type names (Analyst, Craftsman), never codes. Only English is available for now; any other language returns English.
| Argument | Type | Description |
|---|---|---|
| language | string | Language code, for example en. Defaults to English. |
{
"type": "object",
"title": "list_personality_typesArguments",
"properties": {
"language": {
"type": "string",
"title": "Language",
"default": "en",
"description": "Language code, for example en. Defaults to English."
}
}
}List the 16 socionics relationship types: name, one-line tagline, category (comfortable, neutral or tense) and the typeza.ai page of each, with the disclaimer to show alongside them. Call it for an overview of how types relate. Duality is the relation for partners and friends; for work, describe pairs as business relations. Typeza describes patterns between types, not what to do.
No arguments.
{
"type": "object",
"title": "list_relationship_typesArguments",
"properties": {}
}Start the Typeza personality test in this conversation: 41 short questions, no account needed. Returns draft_id and key (keep both and pass them back unchanged), total, answered and the first batch of up to 7 questions. Each question has two statements: ask the person which one fits them better and wait for their answer. Never answer for them and do not show the codes. Then call answer_test with the batch's answers. `language` serves the statements in that language when a full translation exists, otherwise in English. The result is a Typeza type name, never a code.
| Argument | Type | Description |
|---|---|---|
| language | string | Language code, for example en. Defaults to English. |
{
"type": "object",
"title": "start_testArguments",
"properties": {
"language": {
"type": "string",
"title": "Language",
"default": "en",
"description": "Language code, for example en. Defaults to English."
}
}
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 16, 2026.