What is Skymel?
Skymel, from Skymel Inc., sells investigations by an agent called Mel. You describe a goal in plain words, such as a market, a set of prospects or a technology, and Mel searches the public web for small signals: a line in a job post, a footnote in a filing, a comment, a speaker listing, a procurement notice or a code commit. It works out which people and companies each signal refers to and links them into a knowledge graph, with the date, sentence and URL behind every link.
From that graph it delivers an intelligence report, a custom dataset or ongoing monitoring. The report separates what is known, what Mel infers, what points the other way and what is still unclear. Sources named on the site include LinkedIn, X, Reddit, YouTube, GitHub, Hacker News, Substack, patent filings, regulators, earnings calls and status pages.
One-time investigations start at $150, with no subscription and no integration needed. Monitoring and larger scopes are quoted individually.
Is Skymel worth it?
Some questions have no single page with the answer. Which mid-sized software firms are quietly moving off a hosted AI provider? Where are humanoid robots doing paid work rather than demos? You can find out, but it means weeks of job posts, conference agendas and filings.
Skymel sells that legwork by the question. You write the goal, choose how far back and how hard its agent should search, and pay once. Mel collects the small public clues, ties each one to the right company or person, and writes a report in which every conclusion links back to the sentence and page it came from. It also marks what is only inferred and what argues against its own conclusion, which is more candour than most research tools show.
The price is a simple sum. The basic job, 10 days of history, a focused search and one question, is $150. Each step up multiplies it: 30 days doubles it, 6 months quadruples it, a Deep search is 2.5 times and Exhaustive five times, and two or more linked questions double it again. Six months, Exhaustive and connected comes to $6,000.
The catch: you pay before you see anything, and the terms say fees are non-refundable. There are no published turnaround times, no free sample of your own question and no company address on the site. Mel also builds files on named people from their public comments, likes and follows, and the privacy policy, last updated 1 June 2025, offers those people no opt-out. On your own data, the policy says inputs and outputs are not used to train Skymel's models without your written consent, and that its outside AI providers may not train on them either.
Our take: start with the $150 version of a question you already half know the answer to. If the report finds things you missed and every claim checks out against its link, the larger tiers may be worth it. If it mostly repeats the press, you have learnt that cheaply.
Where it earns its place
A strategy lead at a robotics start-up
Find out which factories are paying for humanoid robots, not just trialling them
A 6-month, Deep, single-question investigation costs $1,500 and returns a report tying company disclosures, filings and job posts to each deployment, with a link behind every claim.
The founder of a school photography company expanding into France and Belgium
Build a list of likely buyers in a new country
Ask Mel for a custom dataset of likely buyers; Skymel quotes PixFirst, a customer that used it to find school photography prospects in France and Belgium. A 30-day Focused run is $300.
A head of customer success at an infrastructure software firm
Spot when a large account starts evaluating a rival
Ongoing monitoring watches public issue threads, job specs and talks for that account and explains each change; it is priced by the number of goals watched and how fast changes must surface, so ask for a quote.
Before you commit
- Setup
- Write your goal on the site, pick history, depth and number of questions, pay through Stripe, then write a fuller brief of the company, market or person to investigate; no integration is needed.
- Integrations
- Reports for people; for systems, the docs say delivery may include a REST API, webhooks, structured feeds and data exports, plus connectors listed on its integrations page.
- Data and privacy
- Skymel Inc. says customer inputs and outputs are not used to train Skymel's models without explicit written consent; outside AI providers are chosen that do not permit training on customer data. Deletion within 30 days of a verified request; encryption in transit and at rest. No certifications are published. Which AI tools train on your data?
- Cost at scale
- Priced tiers run from $150 to $6,000 per one-time investigation; monitoring is quoted by goals watched, signal volume, frequency and delivery.
- Limits
- Public sources only, so answers can be incomplete, outdated or wrongly attributed, as Skymel itself states; histories over 6 months need a custom quote; no turnaround time is published.
What’s good
- Every conclusion links to the source sentence and URL
- Separates fact, inference, counter-evidence and gaps
- Clear price formula from $150, no subscription
- Inputs and outputs not used for training without written consent
What’s not
- Fees are non-refundable and paid up front
- No published turnaround time or company address
- Profiles named people from public activity with no opt-out for them
- Monitoring and long histories need a quote
Best for
Founders, strategists and sales leaders who need a sourced answer to one specific market question and would rather pay per question than for a research subscription.
Skymel alternatives
Skymel: common questions
- Is Skymel worth it?
- Skymel is a pay-per-question analyst for markets where the answer is scattered across a hundred small public clues, and you pay before you see any of its work. We rate it 3.3 out of 5.
- How much does Skymel cost?
- One-time investigations from $150, no subscription. The price multiplies by history searched (10 days x1, 30 days x2, 6 months x4), depth (Focused x1, Deep x2.5, Exhaustive x5) and objective (Single x1, Connected x2), up to $6,000; 1 to 3 years of history, strategic programmes and monitoring are quoted. Fees are non-refundable.
- What are the downsides of Skymel?
- Fees are non-refundable and paid up front; no published turnaround time or company address; profiles named people from public activity with no opt-out for them; monitoring and long histories need a quote.
- Who is Skymel best for?
- Founders, strategists and sales leaders who need a sourced answer to one specific market question and would rather pay per question than for a research subscription.
- How much does Skymel cost for a team or in production?
- Priced tiers run from $150 to $6,000 per one-time investigation; monitoring is quoted by goals watched, signal volume, frequency and delivery.
- Is Skymel safe for company data?
- Skymel Inc. says customer inputs and outputs are not used to train Skymel's models without explicit written consent; outside AI providers are chosen that do not permit training on customer data. Deletion within 30 days of a verified request; encryption in transit and at rest. No certifications are published.
- What are the best Skymel alternatives?
- The closest alternatives we have reviewed are Perplexity, Elicit and Clay.
Answers checked Nov 11, 2026.