Dreamwork

Dreamwork is an AI job-search agent for tech roles that ranks live listings against a job seeker's resume, writes a tailored resume, cover letter and screening answers for each role, and submits them on employer application systems such as Greenhouse, Lever and Workday, one tap at a time or through an opt-in Autopilot.

HireAIScore rates Dreamwork 57 out of 100 (grade F, Substantial gaps) under rubric js-v1.0, last reviewed November 6, 2026. Dreamwork ranks 3 of 12 vendors rated in Auto-apply tools, against a category average of 48. That score is drawn from 47 evidence items across 7 rubric criteria for Dreamwork — 42 with a cited source, 5 recording that nothing was located.

Score breakdown

Dreamwork, criterion by criterion

Each rubric criterion is scored out of 100 on the cited evidence, then weighted into the total.

Weighted contribution shown to the right of each bar.

  1. Application Honesty

    Weight 20% · avg 50

    62

    +12.4

  2. Data & Credential Safety

    Weight 20% · avg 50

    54

    +10.8

  3. User Control & Transparency

    Weight 15% · avg 52

    58

    +8.7

  4. Employer & ATS Legitimacy

    Weight 15% · avg 45

    52

    +7.8

  5. Effectiveness Evidence

    Weight 12% · avg 37

    48

    +5.8

  6. Pricing & Cancellation Fairness

    Weight 10% · avg 48

    64

    +6.4

  7. Support & Accountability

    Weight 8% · avg 55

    64

    +5.1

Total weighted score

57.0(rounded to 57 for display)

Category avg is the mean raw score on that criterion across the 12 Auto-apply tools vendors in scope of this rubric, this one included.

Strongest criterion

Support & Accountability

64raw / 100

Contributes 5.1 to the total. 64 against a 55 category average.

Weakest criterion

Effectiveness Evidence

48raw / 100

Contributes 5.8 to the total. 48 against a 37 category average.

Cited evidence

The evidence behind the score

Every score is backed by at least one cited piece of evidence. Open a criterion to see what was found, when, and where.

Download diligence record

Evidence ledger

Items
47
Documentation
34
Public statement
8
Absence
5
With a source URL
42 of 47
Source hostnames
5
Fewest items
5
Data & Credential Safety

These figures measure how thoroughly Dreamwork was reviewed, not how Dreamwork performed — an absence row, recording that nothing was located, is counted like any other item.

Application Honesty

7 items
62
  • DocumentationCaptured Nov 6, 2026

    The About FAQ answers 'Does Dreamwork make things up on my resume?' with 'No', saying each resume and cover letter is built from the user's existing experience and 'changes emphasis and wording for the role, not the facts'.

    https://www.dreamworkhq.com/about

  • DocumentationCaptured Nov 6, 2026

    The July 3, 2026 changelog says 'every rewrite passes a no-fabrication check', that a question 'only you can answer' triggers an in-app banner and email to the user, and that demographic (EEO) questions default to 'prefer not to answer' unless the user opts in.

    https://www.dreamworkhq.com/blog/changelog-2026-07-03

  • DocumentationCaptured Nov 6, 2026

    The June 5, 2026 changelog says the resume optimizer 'stopped inventing' and 'no longer invents a title or headline you never had', which shows the problem existed before that date.

    https://www.dreamworkhq.com/blog/changelog-2026-06-05

  • DocumentationCaptured Nov 6, 2026

    The agent API docs say that if Dreamwork needs more context, the apply workflow 'returns a review state instead of silently proceeding', and the apply action holds until the person confirms.

    https://www.dreamworkhq.com/agents/docs/applications

  • DocumentationCaptured Nov 6, 2026

    Terms section 05 says generated resumes, cover letters and answers 'may contain errors, omissions, fabricated details, biased language, or content that misrepresents your actual experience' and makes the user solely responsible for reviewing all of it before submitting.

    https://www.dreamworkhq.com/terms

  • DocumentationCaptured Nov 6, 2026

    The safety page says that 'By default, Apply sends the application as soon as it's written', and that the user must set the Apply button to Review first to read and edit an application before it is sent.

    https://www.dreamworkhq.com/is-ai-auto-apply-safe

  • Public statementCaptured Nov 6, 2026

    A Dreamwork company post on LinkedIn says that after a September 2026 rebuild, cover letters 'rated ready to send went from 16% to 91%' in blind reviews of real applications. No method is published.

    https://www.linkedin.com/feed/update/urn:li:activity:7509271307453280256/

Data & Credential Safety

5 items
54
  • DocumentationCaptured Nov 6, 2026

    The security page states TLS for all traffic, resume files encrypted at rest, hashed API keys and production access limited to a few engineers. It names Google, Anthropic, OpenAI, Resend, Cloudflare, Railway, Neon and Stripe, says resume content is not used to train general-purpose models, and says in-product deletion removes personal data within 30 days.

    https://www.dreamworkhq.com/security-and-data

  • DocumentationCaptured Nov 6, 2026

    Terms section 05 says that by creating an account, users agree Dreamwork may introduce them to recruiters and share their 'name, contact details, resume'. Dreamwork 'may receive compensation' for introductions, stopping them requires contacting support, and stopping does not retrieve data already shared.

    https://www.dreamworkhq.com/terms

  • DocumentationCaptured Nov 6, 2026

    The Privacy Policy says recruiter introductions 'may be commercial' and may be treated as a sale or sharing under California law, with opt-out only by emailing 'Do Not Sell or Share'. It also lists unnamed 'Application-fulfillment service providers' that receive application answers, including saved self-identification answers.

    https://www.dreamworkhq.com/privacy

  • DocumentationCaptured Nov 6, 2026

    The July 3, 2026 changelog says that when the agent 'creates a job-portal account for you, the credentials stay masked in your inbox until you reveal them', so Dreamwork holds employer-portal credentials it created.

    https://www.dreamworkhq.com/blog/changelog-2026-07-03

  • AbsenceCaptured Nov 6, 2026

    We found no SOC 2, ISO 27001 or penetration-test claim, and neither the Privacy Policy nor the Security and data page mentions storing employer-portal account credentials or how they are protected.

    Have a URL that disputes this?

User Control & Transparency

6 items
58
  • DocumentationCaptured Nov 6, 2026

    The safety page says Apply sends 'as soon as it's written' by default, Review first holds each application for editing, Autopilot is opt-in with caps of 50 a month on Pro and 300 on Dreamer and can be paused, and MCP applies wait for a one-time token that expires after two minutes.

    https://www.dreamworkhq.com/is-ai-auto-apply-safe

  • DocumentationCaptured Nov 6, 2026

    The security FAQ says 'Nothing is sent until you tap Apply on a role, unless you turn on Autopilot', and that Review first makes each application wait for the user to read, edit and send it.

    https://www.dreamworkhq.com/security-and-data

  • DocumentationCaptured Nov 6, 2026

    The September 18, 2026 changelog launching Autopilot says 'It's off until you enable it', can be paused, and shows today's activity, monthly usage and the next scheduled run.

    https://www.dreamworkhq.com/blog/changelog-2026-09-18

  • DocumentationCaptured Nov 6, 2026

    The Agents FAQ says applying, replying to recruiters and changing Autopilot are held until the assistant confirms with a one-time token, and that in Claude and ChatGPT 'applying stays off until you turn it on'.

    https://www.dreamworkhq.com/agents

  • Public statementCaptured Nov 6, 2026

    Dreamwork's own ranking page (which discloses 'We make Dreamwork') concedes under 'Where Dreamwork loses' that 'Dreamwork sends in one tap unless you choose Review first; Jobscan asks every time'.

    https://www.dreamworkhq.com/best-ai-auto-apply-tools

  • AbsenceCaptured Nov 6, 2026

    We found no statement that employers are told an application was written or submitted by AI. Dreamwork's pages describe its applications as reading like ones the user wrote.

    Have a URL that disputes this?

Employer & ATS Legitimacy

8 items
52
  • DocumentationCaptured Nov 6, 2026

    The safety page says Dreamwork 'doesn't log in to LinkedIn or automate Easy Apply' and applies on company career sites and ATSs (Greenhouse, Ashby, Lever, Workday, SmartRecruiters, iCIMS), citing LinkedIn's User Agreement on bot bans.

    https://www.dreamworkhq.com/is-ai-auto-apply-safe

  • DocumentationCaptured Nov 6, 2026

    Terms section 05 appoints Dreamwork 'your limited agent' to 'create and operate a managed sending email address and identity associated with you, and send from it' to employers and ATSs as a standing authorization 'without separate per-application approval'. The user stays responsible for compliance with 'any recipient's terms'. Section 04 bars applying to jobs you don't intend to pursue or spamming employers.

    https://www.dreamworkhq.com/terms

  • DocumentationCaptured Nov 6, 2026

    The security FAQ says employers reach the user 'through a private Dreamwork relay address, so employers do not see your personal inbox', with messages routed through Cloudflare, stored by Dreamwork and forwarded.

    https://www.dreamworkhq.com/security-and-data

  • DocumentationCaptured Nov 6, 2026

    The July 3, 2026 changelog says that when the agent 'creates a job-portal account for you, the credentials stay masked in your inbox until you reveal them'.

    https://www.dreamworkhq.com/blog/changelog-2026-07-03

  • DocumentationCaptured Nov 6, 2026

    The September 25, 2026 changelog says application records 'show the address an employer has on file' and that 'verification codes you request yourself are forwarded to your personal email so you can sign in to employer portals'.

    https://www.dreamworkhq.com/blog/changelog-2026-09-25

  • Public statementCaptured Nov 6, 2026

    In a September 2026 reply to a 3-star review, Dreamwork wrote 'We do have workday solved around 70% of the time'.

    https://www.trustpilot.com/review/dreamworkhq.com

  • DocumentationCaptured Nov 6, 2026

    The auto-apply page says a Dreamwork application submitted through the employer's system 'is indistinguishable from your best manual one', and that if a site blocks automation the pack is left for a manual send.

    https://www.dreamworkhq.com/auto-apply-to-jobs

  • AbsenceCaptured Nov 6, 2026

    We found no ATS API partnership, no disclosure to employers that an AI agent prepared or submitted an application, and no documentation of how email verification is handled for portal accounts the agent creates. We also found no employer or ATS complaints, though Reddit blocked access and X search requires a login.

    Have a URL that disputes this?

Effectiveness Evidence

7 items
48
  • DocumentationCaptured Nov 6, 2026

    The homepage shows interview-invitation emails under 'real interviews our users got', with names blanked and no source or date range, alongside six first-name testimonials (one credited to 'Reddit user') and the line 'Used by thousands who hate job hunting'.

    https://www.dreamworkhq.com/

  • DocumentationCaptured Nov 6, 2026

    The About page says of its product screenshots: 'Screens show a sample candidate. Listings are real. The employers shown with applications and replies are made up.'

    https://www.dreamworkhq.com/about

  • Public statementCaptured Nov 6, 2026

    CEO Colin Johnson's LinkedIn post reports analysing '~8,000 job applications submitted through Dreamwork' and gives relative interview multipliers, such as 50+ applications being 10x as likely to land an interview as under 10. It gives no base interview rate or method.

    https://www.linkedin.com/posts/colin-h-johnson_we-analyzed-8000-job-applications-submitted-share-7523183431921508352-7cBS/

  • Public statementCaptured Nov 6, 2026

    A Dreamwork company post says users landed twice as many interviews from September 19 to 25 as in the week before, and that four in five interviews since June came in September. No counts are given.

    https://www.linkedin.com/feed/update/urn:li:activity:7509271307453280256/

  • Public statementCaptured Nov 6, 2026

    Trustpilot lists 18 unprompted, unverified reviews (TrustScore 4.4). One reports 'several interviews' with finance firms, and a 3-star Swiss review says only the first 3 of about 100 matches were relevant and the CV was 'like 99% copy' of the master CV.

    https://www.trustpilot.com/review/dreamworkhq.com

  • DocumentationCaptured Nov 6, 2026

    The methodology page says the 'active jobs' count behind the 2.1M+ figure 'includes listings awaiting processing and duplicate reconciliation'.

    https://www.dreamworkhq.com/research/methodology

  • AbsenceCaptured Nov 6, 2026

    We found no published response or interview rate per application, no methodology for the LinkedIn figures, and no third-party validation of outcomes.

    Have a URL that disputes this?

Pricing & Cancellation Fairness

5 items
64
  • DocumentationCaptured Nov 6, 2026

    The Cancel and refunds page says users can cancel anytime in the app (Stripe customer portal) or by email, that access runs to the end of the paid period, that payments are non-refundable except where law requires with no proration, that charges made in error are reviewed if reported within 30 days, and that price increases are announced before renewal.

    https://www.dreamworkhq.com/cancel-and-refunds

  • DocumentationCaptured Nov 6, 2026

    The homepage pricing shows Free at $0, Pro at $39 a month or $99 a quarter and Dreamer at $149 a month or $379 a quarter, with 'Card required. Then $39 a month unless you cancel' printed under the 7-day trial button.

    https://www.dreamworkhq.com/

  • DocumentationCaptured Nov 6, 2026

    Terms section 08 says Pro renews automatically, payments are non-refundable except where required by law, and a user who opens a chargeback instead of contacting Dreamwork first may have their account suspended or terminated.

    https://www.dreamworkhq.com/terms

  • DocumentationCaptured Nov 6, 2026

    The October 9, 2026 changelog says eligible trial subscribers receive an email about two days before their first charge, with the renewal date and price.

    https://www.dreamworkhq.com/blog/changelog-2026-10-09

  • Public statementCaptured Nov 6, 2026

    Among 18 Trustpilot reviews, the only one about billing (a 3-star review from September 2026) says the reviewer signed up with a card and cancelled without problems. No billing complaints appear.

    https://www.trustpilot.com/review/dreamworkhq.com

Support & Accountability

9 items
64
  • DocumentationCaptured Nov 6, 2026

    The About page says Dreamwork is 'a product of Ponder Labs, Inc.' and names Colin Johnson (co-founder, CEO), Ben Adamsky (co-founder, CTO) and lead developer Branislav Djalic. It gives support@dreamworkhq.com, with a reply 'usually' within one business day.

    https://www.dreamworkhq.com/about

  • DocumentationCaptured Nov 6, 2026

    The Terms identify Ponder Labs, Inc. as the operator, apply Delaware law with individual AAA arbitration in Wilmington (30-day opt-out), and list hello@, support@ and legal@ contact addresses.

    https://www.dreamworkhq.com/terms

  • DocumentationCaptured Nov 6, 2026

    The blog carries a dated changelog described as 'A weekly record of what shipped', with 22 posts listed from June 5 to October 30, 2026.

    https://www.dreamworkhq.com/blog

  • DocumentationCaptured Nov 6, 2026

    The agent-API changelog has dated entries from 2026-05-14 to 2026-10-06, covering the MCP package, the hosted endpoint and the confirm-token apply flow.

    https://www.dreamworkhq.com/agents/docs/changelog

  • DocumentationCaptured Nov 6, 2026

    The registry record shows dreamworkhq.com was registered on 2026-04-01, so the Dreamwork brand is about seven months old.

    https://rdap.verisign.com/com/v1/domain/dreamworkhq.com

  • DocumentationCaptured Nov 6, 2026

    A March 2024 archive of weponder.io shows Ponder Labs' earlier product, 'Ponder — Web3 Social Surveys', a Farcaster survey and token-reward app.

    https://web.archive.org/web/20240320103333/https://www.weponder.io/

  • DocumentationCaptured Nov 6, 2026

    A February 2026 archive of weponder.io, published by Ponder Labs, shows a different product: daily polls where users 'predict what the world is thinking' and win real money. Dreamwork's job-search product appears only after that.

    https://web.archive.org/web/20260205064519/https://www.weponder.io/

  • Public statementCaptured Nov 6, 2026

    Dreamwork claimed its Trustpilot profile in August 2026 and replied to a critical 3-star review. Reviewers describe the team as responsive and say suggestions were implemented.

    https://www.trustpilot.com/review/dreamworkhq.com

  • AbsenceCaptured Nov 6, 2026

    We found no street address, phone number or public status page: status.dreamworkhq.com has no DNS record, and no Statuspage.io, Instatus or Better Stack page exists under dreamwork or dreamworkhq.

    Have a URL that disputes this?

Editorial notes

Company overview

Dreamwork is an AI job-search agent for tech, AI and startup roles, launched in 2026 by Ponder Labs, Inc., whose earlier products at weponder.io were a web3 survey app and a prediction-poll game. It ranks live listings against an uploaded resume and writes a tailored resume, cover letter and screening answers for each role the user picks. On a paid plan its agent then fills in and submits the employer's own form on Greenhouse, Ashby, Lever, Workday, SmartRecruiters or iCIMS. It does not log in to LinkedIn or automate Easy Apply. By default, tapping Apply sends the application as soon as it is written. 'Review first' is an opt-in setting that holds each one for editing. Autopilot, also opt-in, applies without per-application review to matches above a minimum score, within plan caps. The Free plan writes packs for 5 jobs a day that the user sends themselves. Pro costs $39 a month (50 applications) and Dreamer $149 a month (300). There is also an MCP server that lets Claude, ChatGPT and other assistants run it, a Chrome autofill extension, and a recruiter inbox that tracks each application through the pipeline.

Regulatory exposure

Dreamwork does not ask for LinkedIn or job-board passwords. It documents TLS, encryption of resume files at rest, hashed API keys, named AI and infrastructure processors, a no-training commitment and in-product deletion, but it claims no SOC 2 or other audit. Two data practices cut against users. First, by creating an account users agree that Dreamwork may introduce them to recruiters and share their name, contact details and resume, possibly for compensation. Opting out means contacting support, and data already shared cannot be recalled. Second, the agent creates job-portal accounts for users, and Dreamwork keeps those credentials, but the privacy policy does not mention this. Employers reach the candidate through a Dreamwork-managed relay address, and the Terms authorise a managed sending identity that applies without per-application approval. That resembles the email-alias registration pattern seen at Sonara, and nothing tells employers that an AI agent prepared or submitted the application. On honesty, Dreamwork documents no-fabrication checks and sends questions only the user can answer back to the user. However, its Terms warn that outputs may contain 'fabricated details' and make the user responsible for reviewing everything, while the default Apply mode sends before the user has seen it. Disputes go to individual AAA arbitration in Delaware, with a 30-day opt-out.

Path to a higher score

Making 'Review first' the default for Apply would close the main gap between the Terms, which require users to review every output, and the product, which sends unseen by default. Dreamwork should make recruiter introductions opt-in and add a self-serve off switch. It should also disclose in the Privacy Policy and on the security page that it creates employer-portal accounts and stores their credentials, explain how those credentials are protected and how verification emails are handled, and name its application-fulfillment providers. Telling employers when an application was prepared or submitted by an AI agent would improve legitimacy. Publishing its no-fabrication checks with test results, and a base interview rate with method and date range in place of LinkedIn multipliers and an unsourced inbox gallery, would strengthen honesty and effectiveness scores. An independent security assessment and a public status page would round out accountability.

Risk frame

What applies to auto-apply tools.

Highest job-seeker risk: mass auto-submission can violate job-board / ATS terms and get accounts flagged, and blind applying can send fabricated or mistargeted material under the seeker's name. Credential handling and a review-before-submit step matter most here.

Ranking

Others rated in Auto-apply tools

Ranked 3 of 12 by weighted total under rubric js-v1.0. The ordering is arithmetic on the rubric and carries no view on which tool suits a given hiring process.

  1. #1LandEarly65D
  2. #2Jobscan63D
  3. #3DreamworkThis profile57F
  4. #4Simplify (Simplify Jobs)56F
  5. #5Jobright.ai49F
All auto-apply vendors

Conflicts of interest

No vendor pays for placement, scoring, or removal. Casework - the consulting firm that operates this directory - provides paid services to some vendors. Any active or recent (within 24 months) commercial relationship is disclosed on the affected vendor profile and the review is reassigned to an independent reviewer. See the full policy on About.

Casework has no commercial relationship with this vendor.