Buyer Intent Radar

Find public intent signals from people or companies likely to care about a specific product. Terraform Guard is the first target product, but the radar is now market-agnostic: change the product, market description, keywords, intent terms, and sources without changing code.

• technical sources such as GitHub repositories, GitHub issues, and Hacker News
• known company domains, direct public page URLs, and RSS/Atom feeds
• manually supplied evidence snippets from any allowed source
• configurable positive intent terms and negative-noise filters

The Actor returns ranked leads with evidence URLs, intent scores, proof packs, recommended angles, and human-approved outreach drafts. It also adds a market intelligence brief with buyer moments, source quality, offer experiments, next actions, and a closed-loop learning path.

It is a signal-to-lead intelligence workflow, not just a scraper. Scraping only collects pages. This Actor turns public evidence into a ranked commercial signal: who to inspect, why now, what proof offer to test, which evidence would close the loop, and what repeated product gap should become the next feature or Actor.

Why this exists

Terraform Guard is already published on Apify, but discovery is the missing piece. This Actor is the demand engine: find teams with public evidence of Terraform, platform engineering, CI, cloud security, and AI-agent workflow needs, then produce a clean lead list you can review before outreach.

The same engine can scan another field, for example restaurants, recruiting, real estate, e-commerce, healthcare ops, or SaaS competitors, as long as the user provides the right keywords and public sources.

Local usage

pip install -e ".[dev]"
devops-buyer-intent-radar --max-results 10

With manual evidence:

$devops-buyer-intent-radar --manual-signal "ExampleCo|https://example.com/jobs|Hiring platform engineers for Terraform and CI/CD"

Generate a visual report, Excel-openable CSV, and SQLite database:

devops-buyer-intent-radar \
  --source github_repositories \
  --keyword terraform \
  --keyword opentofu \
  --keyword atlantis \
  --max-results 25 \
  --min-score 35 \
  --report-dir outputs/terraform-guard

Preview mode caps output to 3 leads and skips this Actor's pay-per-event charges. Platform usage costs may still apply depending on Apify billing settings:

$devops-buyer-intent-radar --preview-mode --source github_repositories --keyword terraform

By default the Actor hides weak proof packs and high disqualification risks. Use research mode only when you want broader market exploration instead of a sales-ready shortlist:

$devops-buyer-intent-radar --include-weak-signals --source github_issues --keyword terraform

Use it for Kubernetes or another market by changing keywords:

devops-buyer-intent-radar \
  --product-name "Kubernetes Visual Trainer" \
  --market "Interactive Kubernetes and CKA training workspace." \
  --keyword kubernetes \
  --keyword helm \
  --keyword cka \
  --keyword platform-engineering \
  --report-dir outputs/kubernetes-trainer

Use it outside DevOps by adding market-specific intent terms and generic public sources:

devops-buyer-intent-radar \
  --product-name "Restaurant Booking Radar" \
  --product-url "https://example.com/booking-radar" \
  --market "Restaurants with waitlist, reservation, and booking workflow pain." \
  --keyword restaurant \
  --keyword waitlist \
  --keyword reservation \
  --intent-term "looking for" \
  --intent-term integration \
  --intent-term "manual booking" \
  --source manual_signals \
  --manual-signal "Bistro Ops|https://bistro.example/jobs|Restaurant team looking for waitlist integration and booking workflow automation."

The autoExpandKeywords option derives extra search and scoring terms from the product name, market, and contact persona. The output includes researchScope so the user can see which sources, keywords, expanded keywords, intent terms, and negative filters were used.

Outputs

• Apify Dataset rows: one row per sales-ready lead by default, exportable from Apify.
• Key-value store OUTPUT: complete JSON summary.
• Key-value store RADAR_REPORT_HTML: visual lead report with Market Intelligence Brief and per-lead Proof Pack.
• Key-value store ENGAGEMENT_DASHBOARD_HTML: local-first workflow dashboard for editing outreach and tracking replies.
• Key-value store RADAR_LEADS_CSV: spreadsheet-friendly lead export with buyer moment, urgency, proof strength, micro-offer, and next action.
• Local CLI: optional report HTML, engagement dashboard HTML, CSV, JSON, and SQLite exports.

The engagement dashboard opens Gmail, Outlook Web, or the local mail client with the edited draft. Replies still arrive in the channel used to send the message; the dashboard stores follow-up state in browser local storage and can export the current engagement tracker to CSV or JSON.

MCP

Run the local MCP server:

$devops-buyer-intent-radar-mcp

MCP tools:

• buyer_intent_radar_run
• buyer_intent_record_event
• buyer_intent_sample_inputs

See docs/MCP_USAGE.md for Codex/Claude Code configuration.

Apify input

{
  "productName": "Terraform Guard",
  "productUrl": "https://apify.com/dori1/terraform-guard",
  "keywords": ["terraform", "opentofu", "atlantis", "platform engineering"],
  "intentTerms": ["security", "compliance", "approval", "guardrail", "automation"],
  "requiredTerms": ["terraform", "opentofu", "terragrunt", "infrastructure-as-code"],
  "negativeKeywords": ["student tutorial"],
  "sources": ["github_repositories", "github_issues", "hackernews"],
  "maxResults": 25,
  "salesReadyOnly": true,
  "includeOutreachDrafts": true
}

Pay-per-event hooks

The Actor is wired for:

• intent-radar-run
• qualified-lead

Keep publishing conservative until real cost is measured from 5 to 10 Apify runs. The current local recommendation is $1.00 per run plus $0.10 per qualified lead, because the result is concrete data in the dataset and includes report, CSV, dashboard, and outreach drafts.

Outreach policy

This Actor drafts messages. It does not auto-send messages, post to social media, or contact prospects. That keeps the workflow compliant and protects the creator identity: every public action should be reviewed and approved by a human.