Tracy AI Interviewer
REMOTE · TRACYINTERVIEW.AI
AI video and audio interviews in a natural two-way conversation. A person decides.
Available components
Last scanned (17 hours ago)
Scores higher than 74% of remote servers
Reachable. Last reached 21 Nov 2026.
Category Communication
How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. How we score → Why this is hard to score →
Endpoint Security92
- The endpoint's TLS certificate is valid, in date, and uses a strong key. View diagnostics → Pass
- Authorisation is enforced on tool calls, advertised via RFC 9728 protected-resource metadata. Discovery is public, which costs nothing: no tool can be invoked without a token. View diagnostics → Pass
- HTTPS is enforced; there's no plaintext access path. View diagnostics → Pass
- HSTS check failed: the Strict-Transport-Security header is absent. See how to fix → View diagnostics → Fail
- DNSSEC check failed: this domain isn't protected by DNSSEC. See how to fix → View diagnostics → Fail
- The authorisation server supports Client ID Metadata Documents, the current MCP client-registration mechanism. View diagnostics → Pass
Transport & Reachability100
- Verified streamable-http transport via a live MCP handshake. View diagnostics → Pass
Schema Quality & AI Usability76
- 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name).Pass
- AI-judged instruction clarity (excellent).Pass
- Context-footprint check failed: tool/resource definitions use about 5467 tokens (~455/item across 12 items; 12 tools + 0 resources), over budget; trim descriptions and params. See how to fix → Fail
- Usage-examples check failed: none of the tools include examples. See how to fix → Fail
Stability & Change Management3
- Stability observed for 1 of 30 days with no destabilising changes; credit accrues until the full window elapses.Partial
Tool Coverage100
- 100% of tools have a non-trivial description (not blank, and not just the tool's name).Pass
- 100% of tool parameters carry a description.Pass
Tool Safety100
- No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.Pass
- We read all 12 captured tool definition(s), and no name or description among them implies an irreversible operation.Pass
- An AI judge read all 13 captured unit(s) of tool text and found none that tries to manipulate the model reading it.Pass
Capabilities100
- Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.Pass
Checks we ran (13)
- TLS certificate
- Authorization
- HTTPS
- HSTS header
- DNSSEC
- Transport
- Schema quality
- Stability
- Tool coverage
- Injection markers
- Destructive annotations
- Judged manipulation
- MCP protocol
Does not apply to this hosted endpoint (2)
- Supply Chain SecurityScored on package channels only. This is a hosted endpoint, so there is nothing to install or inspect.
- Provenance & TransparencyScored on package channels only. This is a hosted endpoint, so there is nothing to install or inspect.
The score does not measure everything. What the score does not measure →
Every change we have recorded for this component, newest first. Security-relevant changes are always shown. ▲ marks a change for the better, ▼ a change for the worse; unmarked changes are neutral.
- 21 Nov 26 +1
- Stability: unverified → 0.03 ▲ functional
- 20 Nov 26 +41
- Transport: unverified → pass ▲ security
- Injection markers: unverified → pass ▲ security
- First check of Judged manipulation: pass security
- Authorization: Authorisation is enforced on tool calls, advertised via RFC 9728 protected-resource metadata. Discovery is public, which costs nothing: no tool can be invoked without a token. security
- MCP protocol: unverified → pass ▲ functional
- Schema quality: unverified → 100 ▲ functional
- Tool coverage: unverified → 100 ▲ functional
- First check of Schema quality: fail functional
- First check of Schema quality: excellent functional
- First check of Destructive annotations: pass functional
- First check of Schema quality: fail functional
- First check of Tool coverage: 100 functional
- 18 Nov 26 +14
- HSTS header: unverified → fail ▼ security
- Transport: fail → unverified ▼ security
- Authorization: unverified → pass ▲ security
- 17 Nov 26 −14
- Transport: unverified → fail ▼ security
- HSTS header: fail → unverified ▼ security
- Authorization: pass → unverified ▼ security
- 13 Nov 26 37
First indexed and scored.
Diagnostic detail from the automated scan of this channel: what the scanner observed at each step, so you can see exactly where a check passed or failed. It is informational only and never changes the trust score.
Captured 21 Nov 2026 · Probed https://tracyinterview.ai/mcp
TLS valid
Negotiated TLS 1.3 with TLS_AES_128_GCM_SHA256 .
| Subject | Issuer | Valid from | Valid until | Key | Signature | Serial |
|---|---|---|---|---|---|---|
| CN=tracyinterview.ai | CN=YR1,O=Let's Encrypt,C=US | 30 Oct 2026 | 28 Jan 2027 | RSA 2048 | SHA256-RSA | 6fdfa09a0108b355dd7a066209b7c5d993c |
| SANs: tracyinterview.ai | ||||||
| CN=YR1,O=Let's Encrypt,C=US (CA) | CN=Root YR,O=ISRG,C=US | 3 Sept 2025 | 2 Sept 2028 | RSA 2048 | SHA256-RSA | a20253f15f2691c05dc1ce13b9bcca4e |
| CN=Root YR,O=ISRG,C=US (CA) | CN=ISRG Root X1,O=Internet Security Research Group,C=US | 13 May 2026 | 2 Sept 2032 | RSA 4096 | SHA256-RSA | f24b6d17f9d9ad7cb1c9fea78782699f |
Background: What to check on a remote MCP endpoint →
DNSSEC insecure
Validation of tracyinterview.ai. — Not signed
| Zone | DS | Keys | Algorithms | Outcome |
|---|---|---|---|---|
| . | Trust anchor | 20326, 38696 | 8, 8 | Verified |
| ai. | DS published | 1814 | 8 | Verified |
| tracyinterview.ai. | No DS | Unsigned (proven) parent-signed NSEC/NSEC3 proves an unsigned delegation |
Authentication Enforced and verified
The endpoint asked for a token and published valid RFC 9728 metadata describing how to get one.
| Result | Enforced and verified |
|---|---|
| Enforced | On tool calls |
| HTTP status | 200 |
WWW-Authenticate challenge Bearer error="invalid_token", error_description="Authentication required for this tool", resource_metadata="https://tracyinterview.ai/.well-known/oauth-protected-resource/mcp", scope="tracy"
Bearer error="invalid_token", error_description="Authentication required for this tool", resource_metadata="https://tracyinterview.ai/.well-known/oauth-protected-resource/mcp", scope="tracy" Protected resource metadata
| Document | https://tracyinterview.ai/.well-known/oauth-protected-resource/mcp |
|---|---|
| Retrieved | Yes |
| Resource | https://tracyinterview.ai/mcp |
| Authorisation server | https://tracyinterview.ai |
Background: How OAuth 2.1 works in the 2026 MCP spec →
Transports 2 probes
| Transport | URL | Outcome | Status | Location |
|---|---|---|---|---|
| streamable-http | https://tracyinterview.ai/mcp | Verified | 200 | |
| http (plaintext) | http://tracyinterview.ai/mcp | HTTPS enforced | 301 | https://tracyinterview.ai/mcp |
The tools this component advertises to a client, with an estimated token cost for each. Expand a tool to see its parameters and schema. The per-tool counts are indicative and are not scored directly; the schema's total context footprint is one signal in Schema Quality & AI Usability. A tool's description is untrusted text the model reads on every call, which is what makes this list a security surface and not just an inventory: how tool poisoning works →
create_interview Create an interview ~874
Create and save an approved AI video or audio interview with Tracy AI Interviewer for candidate screening, research or a practice run. Creation requires the host's approval of the exact draft, including video integritySignals, and its matching confirmation; the public link starts on for up to 20 people for 14 days, and nobody is emailed.
| Name | Type | Req | Description |
|---|---|---|---|
| confirm | boolean | – | Whether the host approved the preview. With a valid confirmation, true means approval and false cancels the action. confirmation is the approval code from Tracy's own preview step. That code contains… |
| confirmation | string | – | An approval code from Tracy's own preview step. It contains no personal data, and is sent only back to Tracy to confirm the host approved the preview. The code is valid only for the previewed details… |
| context | string|null | – | Anything else Tracy should know for this interview: job description, company, product, tone. Include only what bears on the interview, and leave it out when there is nothing to add. Do not paste in u… |
| goal | string | yes | What the host wants to find out, as a topic, e.g. "How candidates approach debugging a production incident" or "Why customers cancelled". Do not phrase it as a decision about the people, such as whet… |
| integritySignals | boolean | – | Whether integrity signals run for this video interview, true or false. Optional: they are on by default, so send false only when the host chooses to turn them off, and ask the host to keep them or tu… |
| intro | string|null | – | The first message Tracy says, such as "Hello, I'm Tracy. I'll ask a few questions about your experience." Left out, the standard first message is used. Tracy always opens with a fixed notice that she… |
| language | string | – | The language the interview is held in: en (English), de (German), es (Spanish), fr (French), id (Indonesian), ja (Japanese), ko (Korean), nl (Dutch), vi (Vietnamese). Send the code, e.g. es for Spani… |
| mode | string | – | How the participant joins: video (the default) or audio, where they need only a microphone and no camera. Send audio when the host asks for an audio interview or one without a camera; otherwise leave… |
| outro | string|null | – | The closing message, such as "Thank you for your time. That was everything I wanted to ask." Left out, the standard closing message is used. |
| participants | string | yes | Who is being interviewed, as a group described in the host's words, e.g. "Backend engineers applying for the platform team" or "Customers who cancelled in the last 3 months". It is an audience, not a… |
| questions | array | yes | The questions in the order Tracy asks them, as the host has reviewed them. |
| ref | string|null | – | The host's own reference for this interview, such as an ATS job id like JOB-1042. |
| result_fields | array | yes | The fields the result reports, as the host has reviewed them. An empty list is allowed. |
| template | string|null | – | Copy from createWith. |
| title | string | yes | The interview's title, such as Senior backend engineer screening. |
No output schema declared.
No examples provided.
get_balance Get the balance ~139
Use this when the host asks how many interview results they have left, what their interview allowance and credit balance is, or why invitations or the public link are on hold. Do not use it for another account's balance, for another service's usage or to transcribe a recording. It returns the host's interview allowance and credit balance. Each finished session that produces a report uses one result from them. Invitations are held and the public link stops taking new people when both run out, and they carry on once credits are added. When the balance is low or used up, the reply explains how this affects new invitations and public links and where credits are added.
Input schema present but exposes no named parameters.
No output schema declared.
No examples provided.
get_interview Get an interview ~224
Use this when the host asks what one of their interviews contains or how it is set up, or before changing it with update_interview. Do not use it for meetings held on another service, for another account's interviews, to list interviews (use list_interviews) or to read what participants said (use get_results). It returns one interview in full: its title, purpose, mode (audio or video), language, version, questions, result fields, first and closing messages, whether it is paused, for a video interview whether its integrity signals are on, and whether the warm-up is on. Its complete lists provide the current setup for update_interview.
| Name | Type | Req | Description |
|---|---|---|---|
| interview_id | string | yes | The interview's id, exactly as list_interviews or create_interview returned it in its structured data. It is for tool calls only: never show it to the host; name the interview by its title. Do not gu… |
No output schema declared.
No examples provided.
get_open_link Get the public link ~222
Use this when the host asks for an interview's public link (anyone with the link can join): its address, whether it is on, how many have joined, the join limit and when it expires. Do not use this to build or guess an address, to find a person's personal interview link (list_participants shows those), or to share a meeting link or calendar invitation. It changes nothing and returns the address exactly as the app gives it; never build one from the interview's id. The state says whether it is taking people right now (Accepting) or why not (Off, Paused, No credits, Full, Expired).
| Name | Type | Req | Description |
|---|---|---|---|
| interview_id | string | yes | The interview's id, exactly as list_interviews or create_interview returned it in its structured data. It is for tool calls only: never show it to the host; name the interview by its title. Do not gu… |
No output schema declared.
No examples provided.
get_results Get a participant's results ~314
Use this when the host asks what a participant said, or wants their result, report or answers from an interview. Do not use it for meetings held on another service, for another account's participants or to transcribe a recording, and not to score, rank or reject participants: Tracy only reports what they said. It returns a participant's interview result: what they said to each question, the result fields, and a note for an applicant tracking system. By default it returns the primary result, the one the web app shows: the latest completed session, otherwise the latest partly done one. A person may have had more than one session; set all_sessions to true to list every session with its state and label instead. Each session says how many questions were answered and skipped, or that the count is unavailable, and its report state: ready, processing, failed, or no speech when the participant said nothing. A report appears shortly after a session ends, so a finished session can briefly have none.
| Name | Type | Req | Description |
|---|---|---|---|
| all_sessions | boolean | – | Optional. Set to true to list every session, newest first, instead of only the primary result, for example when the host asks how many times someone took the interview. Leave it out for the primary r… |
| participant_id | string | yes | The participant's id, exactly as list_participants or invite_participants returned it. Do not guess an id, and do not send the person's email address or name in its place: if you do not have the id,… |
No output schema declared.
No examples provided.
invite_participants Invite participants ~448
Invite people by email to a Tracy AI Interviewer interview, using addresses supplied by the host, or self: true for the host's practice run. Send only after approval of the exact recipient preview and its matching confirmation; turn the public link off before a practice invitation, never look up addresses or schedule meetings, and do not treat held invitations as sent.
| Name | Type | Req | Description |
|---|---|---|---|
| confirm | boolean | – | Whether the host approved the preview. With a valid confirmation, true means approval and false cancels the action. confirmation is the approval code from Tracy's own preview step. That code contains… |
| confirmation | string | – | An approval code from Tracy's own preview step. It contains no personal data, and is sent only back to Tracy to confirm the host approved the preview. The code is valid only for the previewed details… |
| interview_id | string | yes | The interview's id, exactly as list_interviews or create_interview returned it in its structured data. It is for tool calls only: never show it to the host; name the interview by its title. Do not gu… |
| people | array | – | Who to invite: one entry per person, each with an email address the host gave. Example: [{"email": "ana@example.com", "name": "Ana Silva"}]. Leave it out when self is true. |
| ref | string|null | – | The host's own reference for one person, such as an ATS candidate id. Used only when inviting a single person. |
| self | boolean | – | true to invite the address the account signed up with, for a practice run where the host takes the interview themselves: the host is not asked for it and people is left out. The preview names the add… |
No output schema declared.
No examples provided.
list_interviews List interviews ~143
Use this when the host asks which interviews they have, wants to find one by its title, or when another tool needs an interview's id. Do not use it for meetings held on another service, for another account's interviews or to read what participants said: it lists interviews only. It returns the host's interviews, most recently changed first, with how many people are on each, how many finished, and whether the public link is on. An audio interview, where the participant needs only a microphone, is marked Audio; a video one is not marked. It has no search or date filter, so find the interview the host means by its title in the list.
Input schema present but exposes no named parameters.
No output schema declared.
No examples provided.
list_participants List participants ~214
Use this when the host asks who is on an interview, who has been invited, has started or has finished, or when get_results needs a participant's id. Do not use it for meetings held on another service, for another account's interviews, or to score, rank or reject participants: Tracy only reports what they said. It returns the people on an interview, newest first, with each one's status (Held, Invited, Opened, Started, Completed, Expired), email status, how many sessions they have had, and their personal link. A participant's id identifies their results in get_results.
| Name | Type | Req | Description |
|---|---|---|---|
| interview_id | string | yes | The interview's id, exactly as list_interviews or create_interview returned it in its structured data. It is for tool calls only: never show it to the host; name the interview by its title. Do not gu… |
No output schema declared.
No examples provided.
set_host_name Set the host's name ~173
Use this when the host gives the name their invite emails should carry, or when invite_participants reports host_name_needed. Do not use this to guess a name, to save a participant's or anyone else's name, to rename an interview, or to change the host's email address or other account details. Save the host's name: invite emails name the host by it, and replies go to the host's email address. Resolves the missing host name reported by invite_participants as host_name_needed, allowing an invitation preview. It sends nothing, saves only the name, leaves everything else on the account as it is, and needs no confirmation. Saving the same name again changes nothing.
| Name | Type | Req | Description |
|---|---|---|---|
| name | string | yes | The host's name, as the people they invite should see it. Example: Linh Do. |
No output schema declared.
No examples provided.
set_open_link Set the public link ~591
Use this when the host wants to turn an interview's public link on or off, or change its join limit or expiry. Do not use this to invite named people by email (invite_participants does that), to schedule a meeting, send a calendar invitation or share a meeting link, or to make an interview public when the host has not asked for a public link. The public link is a public address (anyone with the link can join): nobody needs an invitation until the limit or expiry is reached (20 people and 14 days unless set here; the expiry is 7, 14 or 30 days). An enablement request without a matching confirmation changes nothing and returns a preview with a confirmation. Enabling the public link publishes the interview to anyone with its address and needs the host's approval of the preview and its matching confirmation. The address is available through get_open_link. A confirmation only works for exactly the same arguments. confirm on its own does nothing. Turning it off, or changing only the limit or expiry, needs no confirmation. When a change is saved, the reply also carries the balance and says what to do when credits are low or used up. The confirmation value matches the preview's exact arguments; a changed value or changed arguments produce a new preview without applying the action.
| Name | Type | Req | Description |
|---|---|---|---|
| confirm | boolean | – | Whether the host approved the preview. With a valid confirmation, true means approval and false cancels the action. confirmation is the approval code from Tracy's own preview step. That code contains… |
| confirmation | string | – | An approval code from Tracy's own preview step. It contains no personal data, and is sent only back to Tracy to confirm the host approved the preview. The code is valid only for the previewed details… |
| enabled | boolean | – | true to turn the public link on, false to turn it off. Example: true. |
| expiry_days | – | – | Days the link works, counted from the first time it was turned on: 7, 14 or 30. 14 unless set. Example: 30. |
| interview_id | string | yes | The interview's id, exactly as list_interviews or create_interview returned it in its structured data. It is for tool calls only: never show it to the host; name the interview by its title. Do not gu… |
| join_cap | integer | – | How many people may join through the link, a whole number of 1 or more. 20 unless set. Example: 50. |
No output schema declared.
No examples provided.
suggest_interview Suggest an interview ~641
Draft the title, questions and result fields for a new AI video or audio interview with Tracy AI Interviewer from who the host wants to interview and what they want to learn. Before creation, review the full draft and the video integritySignals choice with the host; it saves nothing, invites nobody and does not handle existing recordings or score, rank or reject people.
| Name | Type | Req | Description |
|---|---|---|---|
| context | string|null | – | Anything else Tracy should know for this interview: job description, company, product, tone. Include only what bears on the interview, and leave it out when there is nothing to add. Do not paste in u… |
| goal | string | yes | What the host wants to find out, as a topic, e.g. "How candidates approach debugging a production incident" or "Why customers cancelled". Do not phrase it as a decision about the people, such as whet… |
| integritySignals | boolean | – | Whether integrity signals run for this video interview, true or false. Optional: they are on by default, so send false only when the host chooses to turn them off, and ask the host to keep them or tu… |
| keep | array | – | Questions to keep as they are when regenerating, such as [{ "text": "What drew you to this role?" }]. |
| language | string | – | The language the interview is held in: en (English), de (German), es (Spanish), fr (French), id (Indonesian), ja (Japanese), ko (Korean), nl (Dutch), vi (Vietnamese). Send the code, e.g. es for Spani… |
| mode | string | – | How the participant joins: video (the default) or audio, where they need only a microphone and no camera. Send audio when the host asks for an audio interview or one without a camera; otherwise leave… |
| participants | string | yes | Who is being interviewed, as a group described in the host's words, e.g. "Backend engineers applying for the platform team" or "Customers who cancelled in the last 3 months". It is an audience, not a… |
| question_count | integer | – | How many questions the draft should hold in all, 1 to 8, any questions in keep counted. Set it only when the host asks for a number of questions, e.g. 4; otherwise leave it out and the draft has 5 to… |
No output schema declared.
No examples provided.
update_interview Update an interview ~960
Use this when the host wants to change an interview that already exists: its title, purpose, mode, language, first or closing message, questions, result fields, the integrity signals or the warm-up. Do not use it to make a new interview (suggest_interview, then create_interview) or to read the current setup (get_interview), or for an existing meeting or recording, an uploaded file, a contact lookup, scoring or ranking participants, or advice alone. Change an existing interview using its current setup from get_interview and the host's edits. Changes to questions or result_fields replace the complete list. Any edit to the setup (title, purpose, mode, language, intro, outro, questions or result_fields) needs both complete lists, questions and result_fields, because they replace what is saved. title, purpose, mode, language, intro and outro are optional and stay as they are when left out; null or an empty value clears the purpose context, intro or outro. Saving makes a new version of the interview. mode is video or audio: an audio interview needs only a microphone from the participant. integritySignals turns the integrity signals of a video interview on or off and needs no lists; an audio interview has no integrity signals, so asking for one is refused and nothing is changed. warmup turns the warm-up before the first question on or off, for a video or an audio interview, and needs no lists. A new language does not translate the questions: send them in that language with the complete lists.
| Name | Type | Req | Description |
|---|---|---|---|
| integritySignals | boolean | – | Whether the integrity signals run for this video interview, such as false to turn them off. Optional: it stays as it is when left out. It is saved on its own, without questions or result_fields. An a… |
| interview_id | string | yes | The interview's id, exactly as list_interviews or create_interview returned it in its structured data. It is for tool calls only: never show it to the host; name the interview by its title. Do not gu… |
| intro | string|null | – | The first message, such as "Hello, I'm Tracy. I'll ask a few questions about your experience." Null or empty clears it. Tracy always opens with a fixed notice that she is an AI interviewer. It comes… |
| language | string | – | The language the interview is held in: en (English), de (German), es (Spanish), fr (French), id (Indonesian), ja (Japanese), ko (Korean), nl (Dutch), vi (Vietnamese). Send the code, e.g. es for Spani… |
| mode | string | – | How the participant joins: video (the default) or audio, where they need only a microphone and no camera. Send audio when the host asks for an audio interview or one without a camera; otherwise leave… |
| outro | string|null | – | The closing message, such as "Thank you for your time. That was everything I wanted to ask." Null or empty clears it. |
| purpose | object | – | The host's answers, if they change. Both participants and goal are needed. |
| questions | array | – | The complete list of questions, in the order Tracy asks them. It replaces the saved list: a question left out is removed, so send every question that stays again, unchanged ones exactly as get_interv… |
| result_fields | array | – | The complete list of fields the result reports, in the order the result shows them. It replaces the saved list: a field left out is removed, so send every field that stays again, unchanged ones as ge… |
| title | string | – | A new title, such as Senior backend engineer screening. |
| warmup | boolean | – | Whether the warm-up runs before the first question, such as true to turn it on. Optional: it stays as it is when left out. It is saved on its own, without questions or result_fields, and applies to a… |
No output schema declared.
No examples provided.
How do I install the Tracy AI Interviewer MCP server?
Tracy AI Interviewer is a hosted endpoint at https://tracyinterview.ai/mcp, so there is nothing to install locally. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.
remote · tracyinterview.ai
claude mcp add --transport http ai-tracyinterview-tracy-ai-interviewer 'https://tracyinterview.ai/mcp'
What is the Tracy AI Interviewer MCP server?
Tracy AI Interviewer is an MCP server listed in the public MCP registry as ai.tracyinterview/tracy-ai-interviewer. AI video and audio interviews in a natural two-way conversation. A person decides. This page covers its hosted endpoint (https://tracyinterview.ai/mcp).
Is the Tracy AI Interviewer MCP server safe to use?
Tracy AI Interviewer scores 79 out of 100 on VerifyMCP. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.
What tools does the Tracy AI Interviewer MCP server expose?
Tracy AI Interviewer exposes 12 tools: suggest_interview, create_interview, update_interview, list_interviews, get_interview, and 7 more. Their descriptions and schemas cost roughly 4,943 tokens of context every time the server is loaded.
Does the Tracy AI Interviewer MCP server require authentication?
Yes. Tracy AI Interviewer asked us for credentials when we connected, so you will need to authorise it in your MCP client before it can do anything.
Is the Tracy AI Interviewer MCP server still maintained?
Tracy AI Interviewer is still listed as active in the MCP registry. We last reached this channel on 21 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.
What did VerifyMCP test on the Tracy AI Interviewer MCP server?
VerifyMCP ran 13 checks on Tracy AI Interviewer, covering TLS certificate, Authorization, HTTPS, HSTS header and 9 more. Supply Chain Security and Provenance & Transparency do not apply to a hosted endpoint, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.