Skip to content
verify mcp Beta VerifyMCP is currently in beta. If you notice any issues, get in touch and we’ll put it right.

RentAHuman

NPM · RENTAHUMAN-MCP

Hire humans for digital microtasks, research, QA, data evaluation, and physical-world tasks.

Available components

−8 this week 67 Trust /100

Last scanned (15 hours ago)

Scores higher than 38% of npm servers

Category Workflow Automation

Trust breakdown (7 categories)

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. How we score → Why this is hard to score →

Supply Chain Security90
  • No malware found by supply-chain analysis.Pass
  • CVE check failed: a known high-severity CVE affects toml 4.3.0, reached via @effect/cli > toml. A fixed version is available. View diagnostics → Fail
  • No install/post-install scripts declared.Pass
  • 35 of 148 dependencies flagged as unhealthy. View diagnostics → Partial
Provenance & Transparency19
  • Repository check failed: the declared repository URL returned HTTP 404. See how to fix → View diagnostics → Fail
  • Provenance check failed: no build-provenance attestation is published. See how to fix → View diagnostics → Fail
  • Clear OSI-approved license (MIT).Pass
  • Actively maintained (last published 3 days ago).Pass
  • Security-disclosure policy not yet verified: we couldn't inspect the source repository.Unverified
Schema Quality & AI Usability77
  • 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 20535 tokens (~285/item across 72 items; 69 tools + 3 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 Management36
  • Stability check failed: the tool surface changed between 3.14.0 and 5.0.0: 59 tool removals, 11 breaking changes, 28 additions. See how to fix → Fail
Tool Coverage96
  • 100% of tools have a non-trivial description (not blank, and not just the tool's name).Pass
  • 87% of tool parameters carry a description.Partial
  • Structured output schemas are declared (96% of tools); any adoption earns full credit.Pass
Tool Safety100
  • No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.Pass
  • All 8 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation.Pass
  • An AI judge read all 71 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
What we tested

Checks we ran (15)

  • Malware scan
  • Known CVEs
  • Install scripts
  • Dependency health
  • Source repository
  • Provenance
  • License
  • Maintenance
  • Schema quality
  • Stability
  • Tool coverage
  • Injection markers
  • Destructive annotations
  • Judged manipulation
  • MCP protocol

Checks we could not complete (1)

These are gaps on our side, not findings about this package. We only credit what we can confirm, so a category made up entirely of them scores 0.

  • Security disclosure in Provenance & TransparencySecurity-disclosure policy not yet verified: we couldn't inspect the source repository.

Does not apply to this package (2)

  • Endpoint SecurityScored on hosted endpoints only. A package has no live endpoint of its own for us to probe.
  • Transport & ReachabilityScored on hosted endpoints only. A package has no live endpoint of its own for us to probe.

The score does not measure everything. What the score does not measure →

Changelog

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.

  • 18 Nov 26 −8
    • Schema quality: 253 → 285 ▼ functional
    • First check of Tool coverage: 96 functional
    • Destructive annotations: All 8 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation. functional
    • Package version: 4.10.0 → 5.0.0 functional
  • 13 Nov 26 +1
    • Stability: fail → unverified ▼ security
    • Tool safety: pass → unverified ▼ security
    • Capabilities: pass → unverified ▼ functional
    • Tool coverage: 100 → unverified ▼ functional
    • Schema quality: 100 → unverified ▼ functional
    • Package version: 4.9.0 → 4.10.0 functional
  • 12 Nov 26 +15
    • Malware scan: unverified → pass ▲ security
    • Package version: 4.8.0 → 4.9.0 functional
  • 11 Nov 26 −15
    • Malware scan: pass → unverified ▼ security
    • CVE-2023-54404 no longer affects this package ▲ security
    • Package version: 4.7.0 → 4.8.0 functional
  • 9 Nov 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 80 to 83.

  • 7 Nov 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 73 to 76.

  • 6 Nov 26 −3

    No change was recorded against any check on this day. Stability & Change Management went from 92 to 73.

  • 5 Nov 26 0
    • Stability: fail → unverified ▼ security
    • Tool safety: pass → unverified ▼ security
    • Capabilities: pass → unverified ▼ functional
    • Tool coverage: 100 → unverified ▼ functional
    • Schema quality: 100 → unverified ▼ functional
    • Package version: 4.6.0 → 4.7.0 functional
Diagnostics

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 · Analysed npm/rentahuman-mcp@5.0.0

Provenance No attestation

The registry publishes no build provenance for this version, so there is nothing to verify.

Result No attestation
Ecosystem npm

Background: How many MCP packages publish verified provenance →

Vulnerabilities 4 findings
ID CVE Severity Vector Fix available
GHSA-4qx8-w7c3-pj3p CVE-2026-114141 high yes
GHSA-9j65-hcg3-gc64 CVE-2026-114143 high yes
GHSA-9qmh-m7c6-f92p CVE-2026-114144 high CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H yes
GHSA-h895-cgqg-jf52 CVE-2026-114142 high yes

Background: What a vulnerability scan can and cannot prove →

Dependencies 148 packages
Packages resolved 148
Stale 35
Tree resolution Complete

Background: SBOMs and build attestations, explained →

MCP tools · 69 exposed · ~20,252 tokens

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 →

Tool Tokens
list_humans ~313

Find people by skill, name, place and hourly rate, best match first. Keep going with the same filters until a page has no `nextCursor`. Use for: finding someone to do a task, or checking who offers a skill. Not for: getting a task done (create_bounty). Money: US cents. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: list_humans({"skill":"photography","city":"Austin","limit":20}) Errors: only the common ones; each error carries `fix`: what to do next.

NameTypeReqDescription
citystring–Ranks people in this city first; others may follow.
countrystring–Country name or code. Ranks people there first; others may follow.
cursorstring–The nextCursor of the previous page. Leave it out for the first page.
limitinteger–Rows wanted. A page can hold fewer, or none; only a page without nextCursor is the end.
maxHourlyRateinteger–Highest USD hourly rate in whole cents; excludes other currencies.
minHourlyRateinteger–Lowest USD hourly rate in whole cents; excludes other currencies.
namestring–Part of their name.
skillstring–Comma-separated skills; matches any of the first 10.
NameTypeReqDescription
dataarray––
nextCursorstring––

No examples provided.

list_messages ~250

Read your conversation’s messages, oldest first. Keep going until there is no nextCursor. Use for: reading a conversation longer than the 100 messages get_conversation returns. Not for: who the conversation is with and whether you can send (get_conversation); finding a conversation (list_conversations). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: list_messages({"conversationId":"c7Rk2mQw9TzL4nXp","cursor":"eyJ2ZXJzaW9uIjoxfQ.5f1c0a9e"}) Errors: only the common ones; each error carries `fix`: what to do next.

NameTypeReqDescription
conversationIdstringyesThe conversation id, from list_conversations or the conversationId of a hire.
cursorstring–The nextCursor of the previous page. Leave it out for the first page.
limitinteger–Rows wanted. A page can hold fewer, or none; only a page without nextCursor is the end.
NameTypeReqDescription
dataarray––
nextCursorstring––

No examples provided.

list_preferred_humans ~146

Everyone you marked as preferred, newest first, with your note. Use for: finding people you liked working with before you post or hire. Not for: taking someone off the list (unprefer_human); people you blocked (list_blocked_humans). Needs: an API key. Hosted sign-in cannot call this yet. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: list_preferred_humans({}) Errors: account_setup_required; each error carries `fix`: what to do next.

Input schema present but exposes no named parameters.

NameTypeReqDescription
dataarray–Everyone you prefer, newest first, up to 500.

No examples provided.

list_qa_runs ~154

List your QA runs, newest first. Send templateId to filter by template. Use for: finding a QA run or its report status. Not for: reading one run’s findings and what each tester sent (get_qa_run). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: list_qa_runs({"templateId":"qa_tpl_4Rk9xW2"}) Errors: only the common ones; each error carries `fix`: what to do next.

NameTypeReqDescription
templateIdstring–The template id the QA template was created with.
NameTypeReqDescription
dataarray–Up to 100, without paging. Some newer runs may be missing.

No examples provided.

list_reviews ~220

Read the reviews of one person, newest first, with who wrote each and whether the job was paid. An unknown id gives an empty list. Use for: checking someone’s track record before you hire them. Not for: reading their profile (get_human). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: list_reviews({"humanId":"h2Lw9QxT4mZp7Rk","limit":50}) Errors: only the common ones; each error carries `fix`: what to do next.

NameTypeReqDescription
cursorstring–The nextCursor of the previous page. Leave it out for the first page.
humanIdstringyesThe id from list_humans. A username does not work here.
limitinteger–Rows wanted. A page can hold fewer, or none; only a page without nextCursor is the end.
NameTypeReqDescription
dataarray––
nextCursorstring––

No examples provided.

list_submissions ~275

List work on one bounty, newest record created first. Continue until there is no nextCursor, even after a short page. Restart to see new work. Use for: finding the work that waits for your review on one bounty. Not for: reading one piece of work in full, with files (get_hire); listing every worker and their money (list_hires). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: list_submissions({"bountyId":"b7Kq2mXw9TzR4nLp","status":"pending_review"}) Errors: only the common ones; each error carries `fix`: what to do next.

NameTypeReqDescription
bountyIdstringyesThe bounty id, or a pasted rentahuman.ai/bounties/... link.
cursorstring–The nextCursor of the previous page. Leave it out for the first page.
limitinteger–Rows wanted. A page can hold fewer, or none; only a page without nextCursor is the end.
statusstring–Only submissions in this status. pending_review is the work that waits for you. Default: every submission.
NameTypeReqDescription
dataarray––
nextCursorstring––

No examples provided.

list_taste_runs ~203

List your preference polls (taste runs), newest first. Continue until there is no nextCursor, even after a short page. New runs shift pages; restart for fresh results. Use for: finding your preference polls and their status. Not for: reading one run’s votes, tally and report (get_taste_run); listing QA runs (list_qa_runs). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: list_taste_runs({"limit":20}) Errors: only the common ones; each error carries `fix`: what to do next.

NameTypeReqDescription
cursorstring–The nextCursor of the previous page. Leave it out for the first page.
limitinteger–Rows wanted. A page can hold fewer, or none; only a page without nextCursor is the end.
NameTypeReqDescription
dataarray–Up to 100 across pages. Some newer runs may be missing.
nextCursorstring––

No examples provided.

list_wallet_transactions ~243

List every movement of money in and out of your wallet, newest first: money added, refunds, bounty funding, and sends paid from the wallet with the worker, their pay and our fee. A send paid by card never moves the wallet, so it is not listed. Keep going until a page has no `nextCursor`. Use for: reading what came in and went out of your wallet, and what each send paid from it cost. Not for: your balance and limits (get_wallet); money held on a hire (list_hires). Money: US cents. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: list_wallet_transactions({"limit":2}) Errors: only the common ones; each error carries `fix`: what to do next.

NameTypeReqDescription
cursorstring–The nextCursor of the previous page. Leave it out for the first page.
limitinteger–Rows wanted. A page can hold fewer, or none; only a page without nextCursor is the end.
NameTypeReqDescription
dataarray––
nextCursorstring––

No examples provided.

list_webhook_deliveries ~307

List queued and sent events for your endpoints, newest first among those read, with each delivery’s status. There is no next page, and with many stored deliveries the newest may be missing: narrow with endpointId, eventType or status. Use for: checking queued and sent events, and finding deliveries that failed. Not for: reading one delivery with its body and tries (get_webhook_delivery); seeing your endpoints (list_webhook_endpoints). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: list_webhook_deliveries({"endpointId":"Wk3pQ8vN1sXr6TzL0aBc","status":"delivered","limit":20}) Errors: only the common ones; each error carries `fix`: what to do next.

NameTypeReqDescription
endpointIdstring–Only deliveries to this endpoint. An id that is not yours gives an empty list.
eventTypestring–Only deliveries of this event type. Leave out for all event types.
limitinteger–Most deliveries wanted. There is no next page.
statusstring–pending: waiting to be sent. claimed: being sent now. delivered: your URL answered 2xx. retry: it failed and will be sent again. failed: no more tries.
NameTypeReqDescription
dataarray––
nextCursorstring––

No examples provided.

list_webhook_endpoints ~162

List every webhook endpoint you registered, newest first, with its events and whether deliveries work. At most 5 exist, so it is one page. The signing secret is never listed. Use for: seeing where your events go, and finding an endpoint that was switched off after failures. Not for: adding an endpoint (create_webhook_endpoint); seeing what was sent to an endpoint (list_webhook_deliveries). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: list_webhook_endpoints({}) Errors: only the common ones; each error carries `fix`: what to do next.

Input schema present but exposes no named parameters.

NameTypeReqDescription
dataarray––
nextCursorstring––

No examples provided.

open_dispute ~408

Hold a hire's pay and ask the platform to settle it, when the work was not done or not as agreed. Use for: work not done, or not as agreed, on a hire whose pay is held. Pay stays held until the platform settles or dismisses the dispute. Use the paying account. Payments in progress can block opening. Listed actions can be refused; hires awaiting worker confirmation may still be disputed. Most accounts get 3 attempts hourly. Needs: an API key. Hosted sign-in cannot call this yet. Money: US cents. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: open_dispute({"hireId":"a3Fh8kQz1WpL0mNc","category":"not_delivered","description":"The worker confirmed the job and never went to the shops. Nothing was sent by the due time.","urls":["https://storage.example/chat-screenshot.png"]}) Repeat: Once this has worked, a repeat does not do it again while that result still stands: it is refused, or answers with what is already there. Errors: no_held_pay, not_disputable, paying_account_required, payment_being_reconciled, dispute_opened_reply_failed, dispute_open_outcome_unknown; each error carries `fix`: what to do next.

NameTypeReqDescription
categorystringyesThe kind of problem. There is no default.
descriptionstringyesWhat went wrong, in your own words: 20 to 2000 characters, not counting spaces at the ends. The platform reads it when it settles the dispute.
hireIdstringyesThe hire id, from list_hires or accept_application.
urlsarray–Up to 10 https links to what shows the problem, each up to 2048 characters. Default: no links.
NameTypeReqDescription
actionsarray–Tools you may call on this hire now. Each tool still checks.
costobject–What this hire costs you as its held money stands now: pay plus service fee. Left out when nothing is held: the fee of a pay-later hire is set when you pay.
disputeobjectyesThe dispute this call opened.
hireIdstringyesThe hire, read again after the dispute was opened: status, payStatus, cost and actions say where it stands. They are returned here also to the owner of a hidden bounty, whom get_hire answers not_foun…
payStatusstringyes–
statusstringyesoffered: waiting for the worker to confirm. active: working, or fixing a redo. submitted: work waits for your review, or the work is done and its pay waits for you to release or pay it. paid: the pay…

No examples provided.

pay_hire ~389

Release the held pay of a hire to the worker. Final. Paying approves the work that waits for review. Use for: paying a worker whose pay is held and whose submitted work you accept. The work is checked once, when the call starts. The reply says what this call released and where the hire stands; get_hire shows the whole hire. Not for: asking for fixes (request_redo); refusing the work (reject_submission). Needs: an API key. Hosted sign-in cannot call this yet. Money: US cents. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: pay_hire({"hireId":"a3Fh8kQz1WpL0mNc","submissionId":"s_7Hd2kQ"}) Repeat: Once this has worked, a repeat does not do it again while that result still stands: it is refused, or answers with what is already there. Errors: not_payable, pay_later_not_supported, hire_kind_not_supported, bonus_decisions_required, bonus_decisions_saved, payment_sent_finishing; each error carries `fix`: what to do next.

NameTypeReqDescription
bonusesarray–Your decision on each bonus of a hire whose bonuses are still to decide. `bonus` in the reply is the bonus money this call released, and get_hire lists the decisions that stand in `submission.review.…
hireIdstringyesThe hire id, from list_hires or accept_application.
submissionIdstring–The submission you pay for: `submission.id` of the hire, from get_hire. When sent, the call is refused if the hire now shows another submission. Default: not checked.
NameTypeReqDescription
actionsarray–Tools you may call on this hire now. Each tool still checks. Left out when there are none, and when status is left out: then read the hire with get_hire.
bonusobject–The bonus money released to the worker with this payment. Left out when it is 0.
hireIdstringyes–
isCreditedToWorkerboolean–What the release reported. true: the worker's balance was credited. Left out (false): the pay left the hold and no credit to the worker's balance was reported yet.
payStatusstring–Left out only when the hire could not be read after this call: then status, payStatus and actions are all left out. Read it with get_hire.
statusstring–Left out only when the hire could not be read after this call: then status, payStatus and actions are all left out. Read it with get_hire.
unusedBonusobject–The bonus money that was not earned. Left out when it is 0.
workerPayobject–The pay this call released for the worker: the hire's pay with the bonus money earned, not counting the service fee. It is in the worker's balance only when isCreditedToWorker is true. Left out only…

No examples provided.

pick_photos ~353

Keep or drop single photos in a photo submission: send keep or drop, one of them per call; to do both, make two calls. An id that is not a photo of this hire is left alone. This sorts photos only; it does not approve the work, refuse it or pay. Use for: choosing the photos you want from a photo submission, before you download them with download_picked_photos. Not for: approving the work and paying (pay_hire); refusing the whole submission (reject_submission). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: pick_photos({"hireId":"a3Fh8kQz1WpL0mNc","keep":["0f3c9a7e21b45d6e8a90","1b7d2e4f6a8c0e2a4b6c"]}) Repeat: A repeat may act again on the thing as it stands then. Read it first. Errors: no_submission, no_photo_submission, not_in_review; each error carries `fix`: what to do next.

NameTypeReqDescription
droparray–Ids of photos to drop, from list_hire_photos. Up to 400. Leave it out or empty when you send keep.
hireIdstringyesThe hire id, from list_hires or accept_application.
keeparray–Ids of photos to keep, from list_hire_photos. Up to 400. Leave it out or empty when you send drop.
NameTypeReqDescription
countsobjectyesEvery photo of the submission, after this call.
updatedintegeryesPhotos of this hire that were marked by this call.

No examples provided.

prefer_human ~289

Mark one person as someone you like to work with. Preferring also takes them off your blocked list. Preferring again only replaces the note. Use for: keeping a list of people you want to hire again. Not for: taking them off the list (unprefer_human); seeing who you prefer (list_preferred_humans). Needs: an API key. Hosted sign-in cannot call this yet. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: prefer_human({"humanId":"h2Lw9QxT4mZp7Rk","note":"Fast, clear photos"}) Repeat: A repeat with the same input, and the same key if you sent one, does not act twice: it returns the same result, finishes the first call, or is refused. Errors: not_found, list_full, account_setup_required; each error carries `fix`: what to do next.

NameTypeReqDescription
humanIdstringyesThe human id, as `id` in list_humans.
notestring–Why you prefer them, for your own list. They never see it. Up to 280 characters. Left out: none, and a repeat without one clears the earlier one.
NameTypeReqDescription
humanIdstringyes–
notestring–Why you prefer them. Missing when there is none.
preferredAtstringyesISO 8601 time with an offset, for example 2026-11-20T09:00:00-08:00.

No examples provided.

reject_application ~239

Turn down one pending application. The applicant is told, with your message if you give one. Use for: saying no to one applicant. The reply says where the application now stands; get_application shows all of it. Not for: ending a hire you already accepted (cancel_hire). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: reject_application({"applicationId":"a3Vn8cQe1RuY6wHs","message":"Thanks, we chose someone closer to the shops."}) Repeat: Once this has worked, a repeat does not do it again while that result still stands: it is refused, or answers with what is already there. Errors: application_not_pending; each error carries `fix`: what to do next.

NameTypeReqDescription
applicationIdstringyesThe application id.
messagestring–What the applicant reads with the rejection. Up to 1000 characters. Default: no message of yours; the applicant then reads a general sentence.
NameTypeReqDescription
actionsarray–Tools you may call on this application now. Each tool still checks.
applicationIdstringyes–
statusstringyes–

No examples provided.

reject_submission ~348

Refuse the submitted work. Final. No money moves at this call: it opens the time in which the worker may appeal. Use for: submitted work that does not meet the bounty and cannot be fixed. `submission.appealDueAt` in the reply shows until when the worker may appeal. The held pay stays held at this call. Not for: fixable work (request_redo); work never submitted (cancel_hire); a problem the platform should settle (open_dispute). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: reject_submission({"hireId":"a3Fh8kQz1WpL0mNc","submissionId":"s_7Hd2kQ","message":"These are photos of another street."}) Repeat: Once this has worked, a repeat does not do it again while that result still stands: it is refused, or answers with what is already there. Errors: no_submission, not_reviewable, hire_kind_not_supported, redo_deadline_pending, rejection_unavailable; each error carries `fix`: what to do next.

NameTypeReqDescription
hireIdstringyesThe hire id, from list_hires or accept_application.
messagestringyesWhat you tell the worker, 1 to 1000 characters.
submissionIdstring–The submission you decide on: `submission.id` of the hire, from get_hire. When sent, the call is refused if the hire now shows another submission. Default: the submission the hire shows now.
NameTypeReqDescription
actionsarray–Tools you may call on this hire now. Each tool still checks. Left out when there are none, and when status is left out: then read the hire with get_hire.
hireIdstringyes–
payStatusstring–Left out only when the hire could not be read after this call: then status, payStatus and actions are all left out. Read it with get_hire.
statusstring–Left out only when the hire could not be read after this call: then status, payStatus and actions are all left out. Read it with get_hire.
submissionobject–The submission of the hire as it stands after this call; get_hire shows all of it. Left out when no submission could be shown: the hire could not be read (status is left out too), or it now shows non…

No examples provided.

report_support_issue ~340

File a report in the support team's queue under your account, with your name and email attached for the team. No email is sent to you. Use for: a platform failure, an unexpected reply, a billing question, or anything you cannot solve yourself. Each call files a new report, so send one once. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: report_support_issue({"summary":"pay_hire answered internal_error twice","description":"I called pay_hire for hire h4Tq9 at 14:02 and 14:05 UTC. Both answered internal_error and get_hire still shows the hire as submitted.","category":"billing","priority":"high","errorCode":"internal_error","sourceTool":"pay_hire"}) Repeat: A repeat may act again on the thing as it stands then. Read it first. Errors: report_refused; each error carries `fix`: what to do next.

NameTypeReqDescription
categorystring–Left out: bug.
descriptionstringyesWhat you did, what you expected, what happened instead. At most 5,000.
errorCodestring–The error code or exception name you got. At most 200.
prioritystring–How urgent it is for you. Left out: none.
sourceToolstring–The tool or workflow where it happened. At most 100.
summarystringyesOne line on what went wrong. At most 160.
NameTypeReqDescription
reportIdstringyesThe id of the report in the support queue.

No examples provided.

request_redo ~394

Ask the worker to fix and resend their work. No money moves. At most 3 times per hire. Use for: submitted work that waits for your review and can be fixed. The worker hands the fixed work in under the same submission id, so check `submission.status` with get_hire before you decide again. Not for: accepting the work (pay_hire); refusing the work for good (reject_submission). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: request_redo({"hireId":"a3Fh8kQz1WpL0mNc","submissionId":"s_7Hd2kQ","message":"Shop 3 is out of focus. Please retake it.","redoDueAt":"2026-11-22T15:00:00.000Z"}) Repeat: Once this has worked, a repeat does not do it again while that result still stands: it is refused, or answers with what is already there. Errors: no_submission, not_reviewable, hire_kind_not_supported, redo_limit_reached; each error carries `fix`: what to do next.

NameTypeReqDescription
hireIdstringyesThe hire id, from list_hires or accept_application.
messagestringyesWhat you tell the worker, 1 to 1000 characters.
redoDueAtstring–When the fixed work is due: 15 minutes to 7 days from now. With no deadline, reject_submission waits for the next hand-in.
submissionIdstring–The submission you decide on: `submission.id` of the hire, from get_hire. When sent, the call is refused if the hire now shows another submission. Default: the submission the hire shows now.
NameTypeReqDescription
actionsarray–Tools you may call on this hire now. Each tool still checks. Left out when there are none, and when status is left out: then read the hire with get_hire.
hireIdstringyes–
payStatusstring–Left out only when the hire could not be read after this call: then status, payStatus and actions are all left out. Read it with get_hire.
statusstring–Left out only when the hire could not be read after this call: then status, payStatus and actions are all left out. Read it with get_hire.
submissionobject–The submission of the hire as it stands after this call; get_hire shows all of it. Left out when no submission could be shown: the hire could not be read (status is left out too), or it now shows non…

No examples provided.

revoke_api_key ~224

Turn one of your API keys off for good. A key can revoke itself; then the next call with it fails. Use for: a key that leaked or is no longer needed. Calls with it are refused from then on. Not for: finding the id of a key (list_api_keys); a replacement key (create_api_key). Needs: an API key. Hosted sign-in cannot call this yet. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: revoke_api_key({"keyId":"Xc81kLm2PqRs9TuvWx0Y"}) Repeat: Once this has worked, a repeat does not do it again while that result still stands: it is refused, or answers with what is already there. Errors: already_revoked; each error carries `fix`: what to do next.

NameTypeReqDescription
keyIdstringyesThe id of the key, from list_api_keys.
NameTypeReqDescription
isRevokedbooleanyesThe key is off for good.
keyIdstringyes–

No examples provided.

search_bounties ~343

Search the open bounties anyone on the site posted, as the public bounty list shows them: featured first, then newest. Use it to find work to apply to, or to see how other people write a bounty (title, description, pay, requirements) before you post your own. A page can hold fewer rows than `limit`, or none; keep going with the same filters until a page has no `nextCursor`. Paging is not guaranteed to reach every match: with a narrow filter it can end before older matching bounties. Not listed: a bounty priced in another currency. Use for: finding open bounties to apply to, or examples to learn from before you post a bounty. Not for: listing the bounties you posted (list_bounties); posting a bounty (create_bounty). Money: US cents. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: search_bounties({"skill":"photography","limit":20}) Errors: only the common ones; each error carries `fix`: what to do next.

NameTypeReqDescription
categorystring–A category slug, such as research or labor-gigs.
cursorstring–The nextCursor of the previous page. Leave it out for the first page.
limitinteger–Rows wanted. A page can hold fewer, or none; only a page without nextCursor is the end.
skillstring–Only bounties with a skill that contains this text.
NameTypeReqDescription
dataarray––
nextCursorstring––

No examples provided.

send_cake ~532

Order a decorated cake delivered to a company, paid from your wallet. Send an idempotencyKey of at most 200 characters. Use for: a cake for a company address. The wallet is charged before the vendor gets the order. The same idempotencyKey answers for the first order as it now stands, ignores new input, and never charges twice. Not for: adding money to your wallet (fund_wallet). Needs: an API key. Hosted sign-in cannot call this yet. Money: US cents. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: send_cake({"recipient":{"company":"Acme Robotics","contactName":"Ada Lovelace","address":"1 Market St","address2":"Floor 4","city":"San Francisco","state":"CA","zip":"94105"},"cakeSize":"8in","design":{"messageText":"Congrats on the launch!"},"cardMessage":"From your friends at Northwind","idempotencyKey":"5f0c7a2e-8d4b-4e1a-9c3f-2b6d1e8a7f40"}) Key: the idempotencyKey in the example is a sample. Replace it with a new unique key for each new call, and keep that key for the call’s retries and any confirmation it asks for. Repeat: A repeat with the same input, and the same key if you sent one, does not act twice: it returns the same result, finishes the first call, or is refused. Errors: wallet_short, spending_limit_reached, not_deliverable, order_unavailable, design_too_large, order_refused, account_setup_required, internal_error; each error carries `fix`: what to do next.

NameTypeReqDescription
cakeShapestring––
cakeSizestringyes–
cardMessagestring–A card that goes with the cake.
designobjectyesExactly one of messageText, imageSvg, imageUrl or imageBase64. A large image goes by imageUrl.
idempotencyKeystringyesA unique string for this call. Repeat it when you retry, so the call runs once.
metadataobject–Your own labels, up to 28 keys, passed to the vendor.
recipientobjectyesA company. Send websiteUrl, or all of address, city, state and zip.
NameTypeReqDescription
chargeobjectyesTaken from your wallet: price + fee.
costobjectyes–
orderobjectyes–

No examples provided.

send_message ~283

Send a text message to the worker in a conversation you are part of. Use for: writing to an applicant or a hired worker. Send an idempotencyKey: the same key with the same conversation and text sends once and answers with that message. Without a key, a repeat sends a second message. Not for: opening chat with an applicant you have not hired (accept_application). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: send_message({"conversationId":"c4Rt7nWq2ZxL9pKs","content":"Hi Dana, the side door is open from 9."}) Repeat: A repeat may act again on the thing as it stands then. Read it first. Errors: conversation_closed, hire_required, message_refused, account_restricted, send_in_progress; each error carries `fix`: what to do next.

NameTypeReqDescription
contentstringyesThe text to send, up to 40000 characters. Text only.
conversationIdstringyesThe conversation id, as on the hire or the application.
idempotencyKeystring–A unique string for this call. Repeat it when you retry, so the call runs once.
NameTypeReqDescription
conversationIdstringyes–
messageobjectyes–

No examples provided.

send_money ~479

Send a tip or a reimbursement to a worker on a bounty you own, after their work on it was released. Paid from your wallet when it covers the total. Send an idempotencyKey. Use for: a tip or an expense after the work is paid. With an idempotencyKey a repeat sends once. Without one, a repeat can send again, and a short wallet gets a card pay link. Not for: adding money to your wallet (fund_wallet); paying for a draft bounty (fund_bounty). Needs: an API key. Hosted sign-in cannot call this yet. Money: US cents. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: send_money({"bountyId":"b7Kq2mXw9TzR4nLp","workerId":"h3Jm8QwLp2Vx","pay":{"amount":1500,"currency":"usd"},"note":"Parking"}) Repeat: A repeat may act again on the thing as it stands then. Read it first. Errors: wallet_short, payment_already_open, work_not_released, amount_over_limit, worker_blocked, worker_cannot_receive, spending_limit_reached, account_restricted, account_setup_required, idempotency_key_reused; each error carries `fix`: what to do next.

NameTypeReqDescription
bountyIdstringyesThe bounty this worker was hired and paid on: a send always belongs to one, as on the website, where money is sent from the chat about that bounty. If you do not know it, list_conversations shows the…
idempotencyKeystring–A unique string for this call. Repeat it when you retry, so the call runs once.
notestring–A memo for the worker, up to 500 characters.
overrideboolean–true sends although you blocked this worker. Default false.
payobjectyesWhat the worker gets, from $0.50 to $10,000. Our fee is added on top.
workerIdstringyesThe worker's profile id, as on their hire.
NameTypeReqDescription
expiresAtstring–When the pay link stops working. Left out when sent.
payUrlstring–The card pay link. Left out when sent.
statusstringyessent: paid from your wallet. waiting_for_payment: nothing was sent yet; pay `transfer.total` at `payUrl` before `expiresAt`.
transferobjectyes–

No examples provided.

set_wallet_controls ~342

Set or turn off your low-balance alert and your two spending caps. A limit you leave out is not changed. Use for: capping what your wallet pays per bounty or per 24 hours, and choosing when you get a low-balance email. Moves no money. Not for: reading your limits and balance (get_wallet); adding money to your wallet (fund_wallet). Money: US cents. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: set_wallet_controls({"spendingCapPer24h":{"amount":50000,"currency":"usd"},"lowBalanceAlertAt":null}) Repeat: A repeat with the same input, and the same key if you sent one, does not act twice: it returns the same result, finishes the first call, or is refused. Errors: account_setup_required; each error carries `fix`: what to do next.

NameTypeReqDescription
lowBalanceAlertAt––You get an email when your balance falls below this. From 0 to $100,000. Null turns it off. Left out: not changed.
spendingCapPer24h––The most your wallet pays in any 24 hours. From 0 to $100,000. Null turns it off. Left out: not changed.
spendingCapPerBounty––The most your wallet funds for one bounty’s hires and prepaid spots; excludes tips and reimbursements. From 0 to $100,000. Null turns it off. Left out: not changed.
NameTypeReqDescription
controlsobject––

No examples provided.

stop_qa_run ~227

Stop a QA run you started. Applicants not yet accepted are turned down; accepted testers are told and may finish. Its schedule keeps going, and no tool stops a schedule. Use for: ending one QA run early. Unused money returns immediately only without assigned work, completed work or paid results; otherwise it returns when the run closes. A stopped run cannot restart. Money: US cents. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: stop_qa_run({"runId":"qr_7Kq2mXw9TzR4nLp"}) Repeat: A repeat with the same input, and the same key if you sent one, does not act twice: it returns the same result, finishes the first call, or is refused. Errors: not_stoppable, run_busy; each error carries `fix`: what to do next.

NameTypeReqDescription
runIdstringyesThe QA run id.
NameTypeReqDescription
refundedSoFarobjectyesA running total of the run’s refunded funding that can grow while the run closes.
runobjectyes–
stopobjectyes–
unsettledTesterCountintegeryesTesters on this run whose work, pay or refund is not settled yet.

No examples provided.

test_webhook_endpoint ~204

Send one signed webhook.test event to one of your endpoints. The reply comes once the event is queued, not once your URL answers. A failed test counts toward switching the endpoint off, like a real event. Use for: checking that your URL receives events and checks their signature, right after creating an endpoint. Not for: reading whether the test arrived (get_webhook_delivery). Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: test_webhook_endpoint({"endpointId":"Wk3pQ8vN1sXr6TzL0aBc"}) Repeat: A repeat may act again on the thing as it stands then. Read it first. Errors: endpoint_disabled; each error carries `fix`: what to do next.

NameTypeReqDescription
endpointIdstringyesThe webhook endpoint id.
NameTypeReqDescription
deliveryIdstringyesRead the outcome with get_webhook_delivery.
eventIdstringyes–

No examples provided.

unblock_human ~217

Take one person off your blocked list, so they can apply to your bounties again. Does not add them to your preferred list. Use for: undoing block_human. Not for: seeing who you blocked (list_blocked_humans). Needs: an API key. Hosted sign-in cannot call this yet. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: unblock_human({"humanId":"h2Lw9QxT4mZp7Rk"}) Repeat: A repeat with the same input, and the same key if you sent one, does not act twice: it returns the same result, finishes the first call, or is refused. Errors: account_setup_required; each error carries `fix`: what to do next.

NameTypeReqDescription
humanIdstringyesThe human id, as `id` in list_humans.
NameTypeReqDescription
humanIdstringyes–
wasBlockedboolean–True when they were on your blocked list and are now off it. False when they were not on it.

No examples provided.

unprefer_human ~204

Take one person off your preferred list. Does not block them. Use for: undoing prefer_human. Not for: seeing who you prefer (list_preferred_humans). Needs: an API key. Hosted sign-in cannot call this yet. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: unprefer_human({"humanId":"h2Lw9QxT4mZp7Rk"}) Repeat: A repeat with the same input, and the same key if you sent one, does not act twice: it returns the same result, finishes the first call, or is refused. Errors: account_setup_required; each error carries `fix`: what to do next.

NameTypeReqDescription
humanIdstringyesThe human id, as `id` in list_humans.
NameTypeReqDescription
humanIdstringyes–
wasPreferredboolean–True when they were on your preferred list and are now off it. False when they were not on it.

No examples provided.

update_bounty ~677

Change, pause or resume your bounty; omitted fields stay unchanged, and null clears only fields that accept null. Lists replace stored lists: copy them whole from get_bounty. Use for: changing a bounty's details or pay. When a wallet charge needs confirmation, resend the same edit with confirmCharge set to the quoted charge. One kind of raise can charge immediately: an eligible sponsored wallet pool, including with other detail edits; charge reports the debit. For card-funded raises that need payment approval, follow the returned payUrl (details.payUrl on a funding_short error). Not for: closing a bounty (cancel_bounty); paying for a draft (fund_bounty). Money: US cents. Replies omit a field's sole empty value. If null is allowed, only null is omitted; false, [] and a present group stay. A present empty group is {}. Field descriptions explain unavailable values. Example: update_bounty({"bountyId":"b7Kq2mXw9TzR4nLp","spots":{"total":4},"confirmCharge":{"amount":2359,"currency":"usd"}}) Repeat: Once this has worked, a repeat does not do it again while that result still stands: it is refused, or answers with what is already there. Errors: charge_confirmation_required, wallet_short, pay_locked, change_pay_alone, locked_after_hire, spots_below_filled, use_cancel_bounty, listing_refused, pay_below_minimum, not_editable, funding_short, bounty_changed, photos_not_available, screening_not_available; each error carries `fix`: what to do next.

NameTypeReqDescription
aiobject––
applicationsobject––
bonusesarray–Replaces all bonuses; [] clears them. Change bonuses separately from pay or spots.
bountyIdstringyesThe bounty id, or a pasted rentahuman.ai/bounties/... link.
categorystring–The kind of work, as in the website’s category picker. Default: unchanged.
checksobject––
confirmChargeobject–The charge of a charge_confirmation_required refusal.
countriesobject––
descriptionstring–a string at most 5000 character(s) long
imageUrlsarray–an array of at most 5 item(s)
locationobject–Replaces the whole location: send every field, null for none.
messagesobject–instructions can change alone. To change any on... message, send all five on... fields, copying unchanged messages from get_bounty.
payobject––
proofobject––
requirementsarray––
schedule–––
skillsarray––
spotsobject––
startsAt––ISO 8601 timestamp with a UTC offset; a changed start must be in the future. Null clears it.
statusstring–paused stops new applications; open resumes them. closed is refused: use cancel_bounty.
timeZonestring–IANA time zone for startsAt and every shift of schedule.
timersobject––
titlestring–a string at most 200 character(s) long
NameTypeReqDescription
actionsarray–Tools you may call on this bounty now. Each tool still checks.
bountyIdstringyes–
changedarray–The top-level fields you sent that this call changed.
chargeobject–What this call took from your wallet; left out when it took nothing.
payUrlstring–A pay raise on a bounty paid by card: the new pay applies once the difference is paid here. Nothing was charged.
refundobject–What this call returned to your wallet; left out when it returned nothing.
statusstringyes–

No examples provided.

Install

How do I install the RentAHuman MCP server?

RentAHuman runs locally as an npm package, launched with npx -y rentahuman-mcp. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

npm · rentahuman-mcp

# add to Claude Code
claude mcp add rentahuman-ai-rentahuman -- npx -y rentahuman-mcp
Common questions

What is the RentAHuman MCP server?

RentAHuman is an MCP server listed in the public MCP registry as io.github.rentahuman-ai/rentahuman. Hire humans for digital microtasks, research, QA, data evaluation, and physical-world tasks. This page covers its npm package (rentahuman-mcp).

Is the RentAHuman MCP server safe to use?

RentAHuman scores 67 out of 100 on VerifyMCP. We recorded 4 known advisories against it as of 21 November 2026. It declares no install or post-install scripts. 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 RentAHuman MCP server expose?

RentAHuman exposes 69 tools: accept_application, block_human, cancel_bounty, cancel_hire, create_api_key, and 64 more. Their descriptions and schemas cost roughly 20,252 tokens of context every time the server is loaded.

Is the RentAHuman MCP server still maintained?

RentAHuman 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 licence is the RentAHuman MCP server under?

RentAHuman declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

What did VerifyMCP test on the RentAHuman MCP server?

VerifyMCP ran 15 checks on RentAHuman, covering Malware scan, Known CVEs, Install scripts, Dependency health and 11 more. One item could not be completed, a gap on our side rather than a finding about RentAHuman: Security disclosure. Endpoint Security and Transport & Reachability do not apply to an npm package, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.