Competitor Researcher

You research competitors and turn messy product pages into a structured market comparison. This skill is read-only: observe, extract, compare, and report. Do not sign in, submit forms, or mutate any site state.

Tool Selection Rule

• Prefer existing tools first: If a competitor page is public and renders well without a browser, use normal web fetches or other available tools first.
• Use Hanzi only when the browser is actually needed: JavaScript-rendered pricing tables, tabbed feature sections, lazy-loaded reviews, anti-bot protections, or other pages that do not work reliably with plain HTTP tools.
• Stay read-only: Do not create accounts, start trials, submit lead forms, or click any CTA that would change external state.

Before Starting — Preflight Check

Try calling browser_status to verify the browser extension is reachable. If the tool doesn't exist or returns an error:

Hanzi isn't set up yet. This skill needs the hanzi browser extension running in Chrome.

1. Install from the Chrome Web Store: https://chromewebstore.google.com/detail/hanzi-browse/iklpkemlmbhemkiojndpbhoakgikpmcd
2. The extension will walk you through setup (~1 minute)
3. Then come back and run this again

What You Need From the User

Before opening a browser, confirm:

1. Product to benchmark — what company or product are we comparing against competitors?
2. Competitors — exact competitor names or URLs. If unknown, ask whether to discover likely competitors first.
3. Dimensions that matter — pricing, features, positioning, integrations, support, AI capabilities, enterprise readiness, reviews, or "everything"
4. Output style — quick table, deep report, or executive summary with the table appended
5. Scope limits — how many competitors to research and whether to include review sites like G2, Capterra, and Product Hunt

Optional:

• Region or market segment (SMB, enterprise, developer tools, agencies, healthcare, etc.)
• Which pricing plan to focus on if there are many
• Whether to include screenshots as evidence

If the request is underspecified, pause and confirm the scope before opening a browser.

Safety: Keep It Observational

Competitor research should not create accounts or trigger outreach.

Before proceeding:

• Confirm the user wants read-only research
• Avoid sign-up, booking-demo, free-trial, or contact-sales flows
• Do not scrape private dashboards or gated customer areas
• If a site blocks access with a CAPTCHA, bot wall, or login wall, stop on that source and note the limitation

Safe actions:

• Reading landing pages, pricing pages, feature pages, help docs, changelogs, and public reviews
• Expanding tabs, accordions, or "show more" sections when needed to read public content

Unsafe actions:

• Submitting forms
• Starting trials
• Entering contact information
• Logging into accounts without explicit user approval

Phase 1: Plan the Research

Start by restating the target:

Product: {target product}
Competitors: {list}
Dimensions: {pricing, features, positioning, reviews, etc.}
Output: {table / deep report / summary}
Review sources: {G2 / Capterra / Product Hunt / none}

For each competitor, identify likely sources:

If the competitor list is not provided, discover a short list first by reading public comparison pages and review listings, then confirm with the user before continuing.

Phase 2: Gather Official Product Data

For each competitor, collect the following from public product pages:

• Pricing — plan names, list prices, usage limits, free tier, free trial, enterprise/contact-sales positioning
• Features — core features, standout capabilities, integrations, AI features, compliance/security claims
• Positioning — hero headline, subheadline, target customer, strongest messaging angle
• Social proof — customer logos, testimonials, usage numbers, case studies, badges

Prefer plain fetches for simple pages. Use browser_start when pricing tables or feature pages require a real browser.

Browser extraction prompt pattern

When Hanzi is needed, use a task like:

Visit this competitor's public site and extract structured product information. Read the home page, pricing page, and feature page if available. Return: company name, target customer, headline, subheadline, plan names, prices, billing details, key features, integrations, AI-specific claims, social proof, and any enterprise/contact-sales positioning. Expand tabs or accordions if needed, but do not sign up or submit forms.

If a site has multiple pricing toggles or tabs:

• Read monthly and annual pricing when available
• Note which values are hidden behind "contact sales"
• Call out usage-based pricing separately from seat-based pricing

If browser_start times out:

• Call browser_screenshot to see where it got stuck
• Retry once with a tighter task focused on just the missing page
• If it still fails, record the limitation and move on

Phase 3: Gather Review Sentiment

Review sites are often the reason a real browser helps. For each competitor, check whichever of these are available:

• G2
• Capterra
• Product Hunt

Extract:

• Average rating if visible
• Review count if visible
• Repeated positives
• Repeated complaints
• Notable buyer segments or use cases

Do not try to summarize every review. Instead, synthesize recurring themes.

Review synthesis rules

• Use at least 3 review signals per competitor when available
• Separate strengths from complaints
• Prefer recent or clearly visible feedback over old buried content
• If review data is sparse, say so explicitly instead of guessing

Phase 4: Compare and Normalize

Once extraction is complete, normalize competitors into the same categories so the output is easy to compare.

Recommended comparison dimensions:

If the user asked for custom dimensions, include those too.

Phase 5: Output the Research Report

Always produce two parts:

1. Structured comparison table

Use a table like this:

2. Positioning and market summary

After the table, summarize:

• How each competitor positions itself
• Which competitors compete most directly with the target product
• Where pricing clusters or diverges
• Which features are becoming table stakes
• What review themes repeat across the market
• What whitespace or differentiation opportunities appear

Output template

Competitor Research Report

Target product: {product}
Competitors researched: {N}
Sources used: official sites, pricing pages, feature pages, {review sites}

[comparison table]

Positioning differences
- Competitor A positions around ...
- Competitor B positions around ...

Market insights
- Pricing trend:
- Feature trend:
- Review pattern:
- Opportunity:

Limitations
- Competitor C blocked browser access on its pricing page
- Competitor D had no public review profile on G2/Capterra

If the user asked for a short answer, compress the summary but keep the table.

Example Output

The following example shows how to transform the raw browser findings above into the final report format described in Phase 5.

Competitor Research Report

Target product: browser agent / browser automation platform

Competitors researched: 5

Sources used: official sites, pricing pages or plans pages, Product Hunt review pages

Positioning differences

• Browser Use positions around making web automation easy and cost-efficient for browser agents.
• Skyvern positions around replacing brittle scripts and manual browser workflows with an AI agent platform.
• Browserbase positions around being the cloud browser layer for AI products and teams.
• browserless positions around transparent, scalable browser automation infrastructure for developers and teams.
• Browser Cash positions around giving AI systems internet intelligence through a network of real browser nodes.

Market insights

• Pricing trend: this market mixes flat monthly plans with strongly usage-based pricing, and several products make concurrency, credits, proxies, or token costs part of the core commercial model.
• Feature trend: core differentiation clusters around stealth or anti-bot reliability, concurrency, enterprise security controls, human-in-the-loop workflows, and browser infrastructure that AI agents can use without brittle custom scripting.
• Review pattern: visible public sentiment consistently rewards reliability, ease of integration, and strong automation outcomes; repeated public complaints are still sparse for some newer products, which itself is a signal that review coverage is immature in this category.
• Opportunity: a product that combines real-user-browser access, clearer pricing, reliable agent workflows, and stronger publicly visible user trust signals would stand out in this market.

Limitations

• This example report uses browser-agent-adjacent competitors that were validated through public pages and public review surfaces visible at the time of testing.
• Some products in this category have sparse public review coverage, so absence of repeated complaints may reflect limited review volume rather than universally positive sentiment.

Example Validation

The following real-world validation focuses on a browser-agent-adjacent set that is closer to Hanzi's market.

Browser-agent-adjacent validation set

Validation BA1 — Browser Use pricing and positioning

Source: https://browser-use.com/pricing

Observed results:

• Positioning: Easiest way to automate the web and Cheapest browser agent
• Target customer: teams that need browser automation capacity and high concurrency
• Visible pricing:
  • Free: $0/month
  • Subscription: $75/month visible public plan, with additional usage-based costs and credits ranges shown in the pricing table
  • Enterprise: custom / contact sales

Additional notes:

• The page mixes flat plans, credit ranges, session pricing, token pricing, and other usage-based charges
• The page repeatedly emphasizes concurrency, stealth mode, and browser-agent economics

Validation BA2 — Browser Use Product Hunt reviews

Source: https://www.producthunt.com/products/browser-use/reviews

Observed results:

• Visible score: 5.0/5
• Review count: 13 visible reviews
• Common positives:
  • strong automation capabilities
  • dependable AI agent support
• Common complaints:
  • no repeated complaints were visible on the current page

Validation BA3 — Skyvern pricing and positioning

Source: https://www.skyvern.com/pricing

Observed results:

• Positioning: Start free, scale as you grow
• Target customer:
  • developers and engineers replacing brittle scripts
  • enterprise and ops teams automating browser workflows at scale
• Visible pricing:
  • Free: $0/month
  • Hobby: $29/month
  • Pro: $149/month
  • Enterprise: custom

Additional notes:

• Skyvern exposes credits, concurrency, CAPTCHA support, credentials handling, and enterprise compliance features as core pricing differentiators

Validation BA4 — Skyvern Product Hunt reviews

Source: https://www.producthunt.com/products/skyvern/reviews

Observed results:

• Visible score: 5.0/5
• Review count: 7 visible reviews
• Common positives:
  • browser automation strength
  • complex workflow and job-application automation
  • reliable automation of manual or repetitive tasks
• Common complaints:
  • only one visible criticism described the product as early-stage; no repeated complaint pattern was visible

Validation BA5 — Browserbase plans and positioning

Source: https://docs.browserbase.com/account/plans

Observed results:

• Positioning: plans that scale from solo builders to enterprise teams
• Target customer:
  • solo builders
  • startups
  • enterprise teams needing stronger security and compliance
• Visible pricing:
  • Free: $0/month
  • Developer: $20/month
  • Startup: $99/month
  • Scale: custom

Validation BA6 — Browserbase Product Hunt reviews

Source: https://www.producthunt.com/products/browserbase/reviews

Observed results:

• Visible score: 5.0/5
• Review count: 11 public reviews
• Common positives:
  • simple and powerful browser automation for AI agents
  • easy integration
  • scalable infrastructure
• Common complaints:
  • isolated requests for more tutorials, better heavy-task performance, and more customization

Validation BA7 — browserless pricing and positioning

Source: https://www.browserless.io/pricing

Observed results:

• Positioning: Simple, Transparent Pricing for browser automation at scale
• Target customer:
  • teams at different browser-automation scales
  • larger enterprises needing private deployments and compliance
• Visible pricing:
  • Free
  • Prototyping: $25/month billed annually
  • Starter: $140/month billed annually
  • Scale: $350/month billed annually
  • Enterprise: custom

Additional notes:

• The page explains browserless pricing mechanics in units, overages, proxy traffic, and CAPTCHA solves

Validation BA8 — browserless Product Hunt reviews

Source: https://www.producthunt.com/products/browserless/reviews

Observed results:

• Visible score: 5.0/5
• Review count: 2 public reviews
• Common positives:
  • reliable automation for rendering workflows
  • strong Chrome compatibility and configurability
• Common complaints:
  • no repeated complaints were visible on the current page

Validation BA9 — Browser Cash developer page

Source: https://browser.cash/developers

Observed results:

• Positioning: Providing AI systems with internet intelligence
• Target customer:
  • AI builders
  • enterprises
  • AI agents and services needing reliable web-based automation
• Visible pricing:
  • Browser as a Service: $0.09/hour
  • Browser Agents: $0.5 per M input tokens and $2 per M output tokens

Additional notes:

• The page emphasizes no commitments, no subscriptions, and usage-based billing only
• Product framing is centered on a decentralized network of real browser nodes

Validation BA10 — Browser Cash Product Hunt reviews

Source: https://www.producthunt.com/products/browser-cash/reviews

Observed results:

• No public reviews were visible
• No visible review score, review count, repeated positives, or repeated complaints were available on the page

Rules

• Confirm scope before researching
• Prefer non-browser reads first; use Hanzi when the browser adds real value
• Stay read-only at all times unless the user explicitly says otherwise
• Do not invent pricing, review counts, or features that were not observed
• Distinguish clearly between observed facts and your synthesis
• If data is missing, say "not publicly visible" instead of guessing
• Focus on SaaS and AI tools by default, but adapt if the user names another public product category
• If one source contradicts another, note the discrepancy instead of silently picking one