For freelancers and small studios
Paste any job post, enquiry or referral. Opportunity Radar scores the fit against your own skills and rates, names the risks hiding in the wording, estimates the real effort, and gives you a price you can defend — in about a minute, before you've written a word.
Fixed price·$3,000–5,000·4 weeks
Unlimited revisions are implied. “Until you're happy” with no round limit — cap it at two in the proposal.
Deliverables are not defined. No mention of theme work, data migration or redirects. Ask before quoting.
The actual cost
It reads fine. The budget looks acceptable. Three weeks in you're doing your fourth revision round on a scope that was never written down, for an effective rate you'd have refused on day one.
Most freelancers triage by instinct and a quick skim. Instinct is good at spotting the obvious disasters and blind to the expensive ones — vague deliverables, an unstated revision policy, a timeline that only works if nothing goes wrong.
Quoting against the client's number instead of your own effort is how a project that should have been $6,000 becomes $3,000 and eleven unpaid hours. The budget tells you what they hoped to pay, not what the work costs you.
Every hour spent writing a proposal for a job you should have skipped is an hour not spent on the one worth winning. Triage is not admin — it is the highest-leverage decision in the week.
How it works
No browser extension, no marketplace integration, nothing to connect. If you can copy the text, it can read it.
Your services, skills and levels, hourly and project rates, the work you want more of, and your own deal-breakers. Everything after this is scored against you — the same post can be an 84 for one freelancer and a 38 for another.
A job post, a client email, a DM, a referral. Add the budget and link if you have them. Budgets in another currency are converted at the day's European Central Bank rate before anything is compared.
A score out of 100 with a clear recommendation, the risks named and explained, the effort in hours, a price built from that effort, and the questions worth asking before you commit.
The score
The model rates each dimension. The server does the arithmetic, with weights that never move. That split is the whole point: identical ratings always produce an identical score, so two opportunities are genuinely comparable — and so are the same opportunity's scores six months apart.
How closely the work maps onto what you actually do, at the level you actually do it. A React project is not a match because it says React; it is a match if the hard part of it is a thing you have done before. This carries a quarter of the score because it is the single best predictor of whether the estimate holds.
Whether the stated budget can support the work at the rates you charge. A budget in euros, pounds or rupees is converted at the day's European Central Bank reference rate before the comparison, and the rate and date are recorded with the analysis so you can see exactly what it compared.
The effective hourly rate once the estimated effort is taken into account, including the parts nobody writes in the post: revision rounds, coordination, the migration of content someone assumed was already done. A $5,000 project at 90 hours is a worse job than a $2,500 project at 18.
A perfect opportunity you have no realistic chance of winning is still an hour you do not get back. This weighs how well you match what they are asking for against how competitive the post looks — so a strong score means both "worth doing" and "worth bidding on".
The difference between "redesign our checkout" and a post that names the pages, the platform and what done looks like. Vague scope is the mechanism behind almost every project that overruns, and it is visible in the text before you accept anything.
Tone, clarity, what they chose to specify, and what they chose to leave out. Dismissive wording, pressure to move off-platform immediately, payment promised later rather than agreed now — these are visible in the post and they predict the relationship.
A job can be well-paid, well-scoped and still the wrong job — more of the work you are trying to do less of. Your stated goals, preferred project types and exclusions all count here, which is how a lucrative maintenance contract can score below a smaller project in the direction you want.
What you get
The score is the start. What makes it worth keeping is everything that happens after you decide to pursue.
Not a vague warning — a specific, catalogued pattern with the words that triggered it. Unpaid test work, contradictory requirements, a timeline that does not fit the work, IP ownership left unsaid.
Built from estimated effort and your own rates, with the reasoning attached — so when the client asks why it costs that, you have an answer better than "that's my rate".
A first draft that uses your writing style and the specifics of the post — plus the questions worth asking before you commit to a number. A starting point you edit, never a send button.
Drag opportunities between stages, record what you decided and why, and log the value, hours and result when work closes. Skipping something is data too.
Which kinds of work you actually win, where your estimates run long, and what your effective rate really is by project type — from the outcomes you recorded, not from a benchmark about someone else.
Set a follow-up date on anything you sent and get a reminder in the browser. Most lost work is not lost to a competitor; it is lost to silence nobody chased.
Export everything as a file, or delete your account and its contents outright, from inside the app. What you paste is never used to train a model, never shown to anyone else and never sold.
Upwork, Freelancer, LinkedIn, email, a DM, a referral from a friend. It reads plain text, so there is nothing to connect and nothing to break when a marketplace changes its layout.
Pricing
Monthly, in US dollars. Cancel whenever — you keep the plan until the period you paid for ends, and your pipeline stays yours either way.
$0forever
Enough to score a month of real opportunities and find out whether it changes your mind.
$19.99/month
For someone bidding regularly who wants the pricing and insight tooling behind every decision.
$49.99/month
For a small studio or an agency-of-one with enough inbound that triage is a daily job.
Every plan starts free. Paid plans are billed monthly and can be cancelled from inside the app at any time.
Questions
A language model rates the seven dimensions above. The server — not the model — combines them using fixed weights and applies the deterministic penalties for your own stated limits.
That split matters. Ask a model for a final score and it gives you a different one on Tuesday. Ask it to rate dimensions and do the arithmetic yourself, and the same ratings always produce the same number, which is what makes two opportunities comparable at all.
Yes — it reads plain text. Upwork, Freelancer, Contra, a LinkedIn message, an email enquiry, a Slack message from a former client. There is nothing to connect and nothing that breaks when a site redesigns.
It is stored in your own workspace and sent to an AI provider to be analysed. It is never used to train a model, never shown to another user and never sold. You can export everything or delete your account and its contents from inside the app.
Job posts often contain someone else's confidential information. Paste only what you are entitled to share — the privacy notice explains this in full.
It writes a first draft in your voice, using your profile and the specifics of the post, along with the questions worth asking the client before you commit to a number.
It is a starting point you edit. A proposal that reads like it was generated is worse than no proposal, and the draft exists to save you the blank page, not the thinking.
It is a decision aid, not a verdict. Where a score lands close to a recommendation boundary the analysis is re-sampled and the median taken, because language models genuinely return different numbers for identical input — and a borderline call is exactly where that variance would change your mind for no good reason.
Every score shows its dimension ratings and the reasoning behind each flag, so you can disagree with it on specifics rather than in general.
Yes. Cancelling stops the renewal and you keep the plan until the period you already paid for ends — no refund clawback, no losing access the moment you decide not to renew.
Afterwards your opportunities, pipeline and outcome history stay available on the free plan. Losing a subscription should not lock you out of your own record.
No. It runs in the browser on desktop, tablet and phone, and it can be added to a home screen like an app. There is no extension and nothing to keep updated.
Ten analyses a month, free, no card. If it changes your mind about one job this month, it has already paid for a year of Pro.
Opportunity Radar