Company Deep Research — SEC, GitHub, DNS & Social
Pricing
from $1,000.00 / 1,000 company researcheds
Generate comprehensive company research reports from 7+ sources: SEC filings, stock data, Wikipedia, GitHub, Trustpilot reviews, DNS records, and social media verification. One domain in, full intelligence report out.
Pricing
from $1,000.00 / 1,000 company researcheds
Rating
0.0
(0)
Developer
Ryan Clinton
Maintained by CommunityActor stats
2
Bookmarked
16
Total users
9
Monthly active users
10 days ago
Last modified
Categories
Share
Record Type
recordType
Optional
Discriminator: 'company-report' for a successful research record, 'error' for an error record
Instant (1-3 word state)
instant
Optional
Compact state read: label (1-3 words: "High Growth", "Wind Down", "Stable", "M&A Active", "Launching", etc.), confidence (0..1), state enum, semantic color (green/yellow/red/blue/grey — UI-layer maps to icons; emoji is opt-in elsewhere). Designed for dashboards / mobile / Slack tiles.
TLDR (paste-ready one-liner)
tldr
Optional
1-sentence + topRisk + topOpportunity + needsAttention. Designed for Slack subjects, email previews, and AI agent tool-call summaries. Read this if you only have 5 seconds.
Why now (single-line trigger)
whyNow
Optional
Explains why THIS run matters right now: trigger (what fired), change (directional shift), importance (relative-to-portfolio framing). Returns null when no notable trigger — better than emitting noise. Built from event severity + correlation confidence + drift + portfolio outlier status.
Story (single canonical narrative)
story
Optional
ONE coherent narrative collapsing tldr + whyNow + deltaStory + changeSummary into four 1-line slots: now / trend / decision / outlook + a stitched 2-3 sentence narrative. Use this when you want a single field to summarise the whole run — for AI agent tool calls, exec emails, daily-digest subjects.
Summary (start here)
summary
Optional
Hero synthesis block — headline, oneLine, keyTakeaways[], whatToCheck[], confidence (with dataCoverage + signalStrength + stability). Built deterministically from the modules below — open this first.
Priorities (THE canonical decision surface)
priorities
Optional
This is the canonical decision surface — events / correlations / anomalies / drift / predictions are supporting evidence below. Top 5 ranked decisions. Each priority has rank, type, severity, headline, reason, whyItMatters, recommendedAction (concrete next step), evidence[], timeToImpact (immediate / days / weeks / months). Most users only need to read priorities[0].
Delta story (7d / 30d / 90d narratives)
deltaStory
Optional
Temporal compression — what changed in the last 7d, 7-30d, and 30-90d windows, plus a stitched narrative paragraph. Coverage enum tells you which windows actually have data. Requires snapshot history.
Watchlist actions (CRM-workflow-ready)
watchlistActions
Optional
CRM-workflow-ready actions that bridge into sales / monitoring / due-diligence pipelines: move-to-active-pipeline, schedule-weekly-monitoring, schedule-daily-monitoring, trigger-outreach, pause-outreach, open-due-diligence-ticket, add-to-deal-pipeline, remove-from-active-list, flag-for-security-review, subscribe-to-status-page, archive-as-dormant. Each has type + label + rationale + confidence.
Decision memory (last action + inferred outcome)
decisionMemory
Optional
Closes the feedback loop. When you pass lastAction input on subsequent runs, the actor compares the current state vs the snapshot at action time and infers outcome (engaged / escalated / no-response / no-change / resolved / too-soon-to-tell) + effectivenessScore + pattern (e.g. "action X repeated 3 times — review if it's still right") + actionTypeFrequency per entity. Outcome is honestly inferred from observable signal changes only.
Decision risk (false-positive vs false-negative cost)
decisionRisk
Optional
Per priorities[0]: falsePositiveCost (cost of acting when you shouldn't) + falseNegativeCost (cost of missing) + reversibility + blastRadius + asymmetry (symmetric / fp-dominated / fn-dominated) + actEvenIfUnsure boolean (true when fn-dominated + reversible + low-blast — bias to action). Lets users answer "should I act EVEN IF confidence is low?"
Time to regret (when does NOT acting become a mistake?)
timeToRegret
Optional
Per priorities[0]: window (e.g. "24-48h" / "7-14 days" / "weeks") + urgencyCurve (very-steep / steep / moderate / gradual / flat) + deadlineHint (approximate ISO date when no longer worth acting) + reason + epistemicStatus (this is heuristic, not a known event). Encodes 'fear of missing the window' explicitly — people act on regret-avoidance more than severity / confidence.
Inaction outcome (what happens if you do NOTHING)
inactionOutcome
Optional
Loss-framing complement to timeToRegret. Per priorities[0]: expectedState (what happens if you don't act) + confidence + timeframe + reason + epistemicStatus. Humans decide on regret AND loss; this completes the pair. Honest disclosure: pattern-based forecast, not measured outcome.
Signal independence (3 signals or 1 echoed 3 times?)
signalIndependence
Optional
Score (0..1) showing whether the events / correlations / anomalies / drift are truly independent or echoes of one underlying change. signalCount + distinctSourceCount + interpretation + warning (fires when score is low). Catches the 'looks like 3 corroborating signals but really 1 underlying delta' trap.
Primary action only (permission to ignore the rest)
primaryActionOnly
Optional
Schema-level flag telling users "do priorities[0] only this run; ignore the rest." Fires only when conditions are unambiguous: single high-severity priority + steep urgency curve + bias-to-act risk profile. Permission to not scroll. Returns enabled:false with explanation when conditions don't warrant simplification.
Decision posture (execute / analyze / hold / monitor)
decisionPosture
Optional
The psychological switch from analysis-mode to execution-mode. Mode enum: execute (4+/5 conditions met — bias-to-act risk + urgency + signal independence + horizon alignment + primaryActionOnly), hold (multiple contradictions / low independence / misaligned horizons / 3+ uncertainty areas), analyze (priority exists but conditions not strongly favouring either), monitor (no actionable priority). Each carries reason + instruction + confidence.
Priority computation (weight transparency)
priorityComputation
Optional
Exposes the weight stack that produced priorities[0]: dominantFactors[] (which signals contributed and at what weight) + suppressedFactors[] (which signals were weighted down and why) + weightStackVersion (stable identifier — bumps on rule changes) + explanation. When users disagree with priority ranking, this is the audit trail.
Time horizon alignment (short-term vs long-term coherence)
timeHorizonAlignment
Optional
Catches "this is urgent AND accelerating" misreads when reality is "short-lived spike inside long-term stability". Status: aligned / misaligned / partial / insufficient-history. Pairs timeToRegret (short-term) against trajectory (long-term) and emits reason + interpretation.
Action guard (saturation protection)
actionGuard
Optional
Tells users what NOT to do. recommendedMaxActions (typically 1-3) + totalActionsAvailable + suppressedActions + reason + rationale. Diminishing returns set in beyond 2-3 actions per run because attention is finite, downstream automation gets noisy, and acting on too many signals erodes signal-to-noise on subsequent runs. The system that tells users to stop is the system they trust.
Identity drift (is this company becoming something else?)
identityDrift
Optional
Tracks transformation, not just activity. Compares the current archetype + lifecycle against the previous portfolio entry; emits from + to + confidence + signals[] + strategic implication. Example: fintech-api → platform suggests engagement strategy should shift from point-solution to platform-partnership pitch. Requires a previous portfolio entry on the same domain.
Why this decision (1-line human rationale)
whyThisDecision
Optional
1-line human-readable rationale for priorities[0]. Compressed version of explain.entries[0] for execs / non-engineers / Slack. Mentions whether the priority is correlation-driven, anomaly-driven, or single-event-driven; whether the counterfactual confirms causality; and whether decision-risk asymmetry suggests biasing to / away from action.
Counterfactual (what without the top signal?)
counterfactual
Optional
Removes the signal driving priorities[0] and recomputes priorities[0] + trajectory + instant.label. Output: droppedSignal + withoutThisSignal block + interpretation. Isolates which signal is load-bearing — useful for sanity-checking that the recommended action is causally tied to the right evidence, not coincidence.
Portfolio pressure (attention share + displacement)
portfolioPressure
Optional
Only when portfolioId set + 4+ entries. relativeUrgency (highest-this-week / top-tier / middle / low) + attentionShare (0..1 fraction of total portfolio alert intensity) + displacement (which lower-priority entries are pushed out of focus) + displacedDomains[] + recommendedFocusShift boolean. Answers "what should I STOP paying attention to?" — the inverse of the standard attention-add framing.
Predictions (forward-looking)
predictions
Optional
Deterministic forward-looking predictions: product-launch-likely / acquisition-imminent / infra-migration-likely / funding-event-likely / rebrand-likely / security-audit-likely / wind-down-likely / platform-expansion-likely. Each carries confidence (0..1), timeframe, evidence, headline, rationale. Built by layering rules over events + anomalies + correlations + trends — no LLM, no external data.
Trajectory (direction + velocity)
trajectory
Optional
Company trajectory: direction (accelerating / steady-growth / stable / decelerating / declining / unknown) + velocity (high / medium / low / none / unknown) + confidence + plain-English explanation. Component deltas exposed for downstream re-weighting. Requires 2+ snapshots.
Change Summary (one-sentence narrative)
changeSummary
Optional
One-sentence narrative of what changed since last run: headline + direction (growth / stable / decline / mixed / no-data) + confidence (high/med/low) + keyEvents[] (top 1-3 evidence items).
Triggers (precomputed booleans for automation)
triggers
Optional
Precomputed booleans for downstream automation routing (Zapier / Make / Slack / agent tool calls). Includes highSeverityEvents, possibleAcquisition, productSignals, infraMigration, emailInfraChange, brandRefresh, communityTraction, securityRiskHigh, rapidGrowth, dormancy, needsHumanReview. Filter with WHERE triggers.X = true instead of parsing prose.
Suggested actions (ready-to-send payloads)
actions
Optional
Ready-to-send action payloads for downstream automation: webhook-payload (generic JSON), crm-enrichment-hubspot (HubSpot Company properties), slack-block-kit (pre-formatted Slack message), jira-issue (high-severity-only), email-digest (subject + body), csv-row (flat one-row representation). Each has type, target, rationale, payload.
Next-actor suggestions (orchestration hints)
nextActors
Optional
Suggested follow-up Apify actors to chain after this one: SEC EDGAR Filing Analyzer (when CIK detected), Website Tech Stack Detector (when in-actor detection is incomplete), Person Enrichment Lookup (when sales / careers / B2B SaaS signals present), Lead Enrichment Pipeline, WHOIS Domain Lookup. Each has actor (full slug), label, reason, suggestedInput (pre-filled).
Playbooks (declarative IF-THEN strategy)
playbooks
Optional
Declarative IF-THEN strategy rules that fire on this run: expansion-phase-engagement, wind-down-de-prioritise, m-and-a-imminent, infra-overhaul-watch, security-soft-target, product-launch-watch, funding-round-watch, rebrand-or-pivot-watch. Each carries triggered conditions, implication, recommendedStrategy (concrete), suggestedCadence.
Portfolio context (rank + percentile vs your portfolio)
portfolioContext
Optional
Only emitted when input portfolioId is set. Compares this entity against the user's other portfolio entries: rank (e.g. "3/120"), percentile (0..1), outlier (boolean — top or bottom 5%), reason (plain-English why), portfolioMedians. Each user's portfolios are isolated by Apify per-user named-store sandboxing.
Portfolio feed (rolling alerts + topMovers + newEntrants)
feed
Optional
Only emitted when input portfolioId is set. Cross-run aggregation across the user's portfolio: rollingAlerts[] (last 30 alerts, capped at 14 days), topMovers[] (entities whose alert intensity changed most), newEntrants[] (domains added to the portfolio in the last 7 days). Designed as a daily / weekly intelligence feed.
Normalized scores (percentile vs portfolio)
normalized
Optional
Only emitted when input portfolioId is set AND portfolio has 4+ entries. Each metric is reported as a percentile (0..1) rank against the rest of the portfolio: technicalMaturityPercentile, securityPosturePercentile, subdomainCountPercentile, githubStarsPercentile, hackerNewsMentionsPercentile. Solves "is 186 repos a lot?" by giving relative context.
Cluster (similar companies in your portfolio)
cluster
Optional
Only emitted when input portfolioId is set AND portfolio has 3+ entries. Membership in a cluster of portfolio companies sharing the same fingerprint / infra signature: id, basis, sizeInPortfolio, similarCompanies[], position (leader/middle/lagger/lone), rationale.
Cold-start guidance (when portfolio < 4 entries)
coldStart
Optional
Only emitted when portfolioId is set AND portfolio has fewer than 4 entries (otherwise null — features unlocked). Provides bootstrap guidance: portfolioSize + needsMore + suggestedSeeds (5 well-known public companies matching this entity's archetype to add as portfolio seeds). Solves the 'new users get a worse product' cold-start problem.
Decision quality (meta trust layer)
decisionQuality
Optional
Meta-layer trust block: completeness (fraction of decision-layer fields populated), consistency (degree to which signals reinforce each other across cross-checks), contradictions[] (detected internal inconsistencies — e.g. "high infra complexity but zero open-source presence"), summary.
Drift signals (pattern changes beyond anomalies)
drift
Optional
Pattern-change detection beyond per-metric anomalies: behavior-change / velocity-shift / composition-shift / attention-shift. Detects e.g. "GitHub repo growth slowed from +2/run to 0/run" or "Subdomain expansion accelerated recently". Requires 5+ snapshots.
Explain (reasoning chain)
explain
Optional
Reasoning-chain exposure for the top decision-layer outputs: each entry has target (which output is being explained), derivedFrom[] (which raw signals fed it), rule (which logic fired), optional weights. Plus principles[] listing the actor's reasoning commitments (deterministic, no LLM, audit-friendly, bounded state, honest abstention).
Company graph (relationships + clustering keys)
graph
Optional
Treat this domain as a node in a network: relatedCompanies[] (suspected sub-brands / acquisitions derived from notable subdomains), sharedInfrastructureKey (cluster with companies on same infra), sharedEmailInfraKey, sharedTrackingKey (companies that share an ad-network footprint), suspectedSubBrands[]. Build company graphs in BI by joining on these keys.
Memory (milestones + patterns from history)
memory
Optional
Cross-run memory: historyDepth (e.g. '47 days across 7 snapshots'), milestones[] (first-occurrence events: first-subdomains-detected, first-github-presence, first-infra-migration, first-email-infra-change, first-brand-refresh), patterns[] (plain-English: 'consistent subdomain growth (5 of last 6 runs)').
Positioning (when compareTo is set)
positioning
Optional
Competitive positioning vs the peer cohort: category (archetype), rank, rankBasis (which metric), leaders[] (peers ahead of us), strengths[] / weaknesses[], summary (one-line). Only emitted when input compareTo is populated.
Uncertainty (where the actor is unsure)
uncertainty
Optional
Honest catalogue of where this report is uncertain: company-name-detection / acquisition-detection / lifecycle-stage / company-type-classification / archetype-classification / tech-stack-detection. Each carries area, reason, our confidence (0..1), and a concrete suggestedFix. Builds trust by surfacing failure modes upfront.
Intelligence
intelligence
Optional
Computed classification: companyType (startup / scaleup / public / enterprise / private), archetype (developer-platform / saas / fintech / ecommerce / etc.), businessModelHints[], technicalMaturityScore + level, openSourceStrength, infraComplexity, operationalMaturity, growthSignals[], riskSignals[], notablePatterns[]. All deterministic — no LLM.
Lifecycle stage
lifecycle
Optional
Company lifecycle stage: nascent / growing / scaling / mature / declining / dormant / unknown. Includes confidence + plain-English signals[]. Derived from age + GitHub growth + careers presence + recent activity.
Scoring (signal weight transparency)
scoring
Optional
Per-score breakdown of which factors fired, their weights, and their actual contribution. Covers technicalMaturity (12 factors), securityPosture (10 factors), operationalMaturity (5 factors). Use this to audit / explain / re-weight scores downstream.
Correlations (compound patterns)
correlations
Optional
Multi-event compound patterns detected across the events[] array. Pattern enum: product-launch / infra-migration / acquisition / pivot / wind-down / security-overhaul / funding-event / rebrand. Each carries confidence (0..1), evidence[], and an explanation.
Anomalies (statistical outliers)
anomalies
Optional
Z-score-based anomaly detection from snapshot history (requires 4+ prior runs). Types: subdomain-spike / subdomain-drop / github-burst / github-stall / hn-spike / sec-filing-cluster. Each carries detail, interpretation, severity, zScore.
Use-case views (sales / security / investor / engineering)
views
Optional
Same data, four audience framings — each with angle (one-line), hooks[] (why this audience cares), risks[] (what might disqualify), nextSteps[] (concrete actions). Enables one report → 4 different consumers.
Events (typed diff classification)
events
Optional
On scheduled / re-run, classifies the raw diff into typed events: CORPORATE_UPDATE / PRODUCT_SIGNAL / INFRA_EXPANSION / INFRA_MIGRATION / INTEGRATION_ADDED / EMAIL_INFRA_CHANGE / BRAND_REFRESH / POSSIBLE_ACQUISITION / COMMUNITY_TRACTION — each with severity (low/medium/high), evidence string, and explanation.
Fingerprint
fingerprint
Optional
Hashes for clustering, dedup, and competitive comparison: techStackHash, infraSignature, orgSignature, securityHeadersHash.
Security Posture
securityPosture
Optional
Composite security score (0..1) from DMARC enforcement, SPF, CAA, security.txt, and HTTP security headers (HSTS, CSP, X-Frame-Options, X-Content-Type-Options, Referrer-Policy, Permissions-Policy). Includes per-control issues[] and strengths[].
Trends (multi-run)
trends
Optional
Multi-run deltas computed from snapshot history (last 10 runs per domain). Includes 30d / 90d deltas for subdomains, GitHub repos / stars, Hacker News mentions, SEC filings — plus infra stability and change frequency.
Gaps (partial-fail intelligence)
gaps
Optional
When a module came up empty: which one, the impact (low/medium/high), and a plain-English reason. Helps users distinguish 'no data exists' from 'actor broke'.
Entity ID
entityId
Optional
Stable cross-system identifier ({domain}|{slug-of-companyName}) — use this as a join key in downstream pipelines.
Output profile
outputProfile
Optional
Which output profile produced this record: 'analyst' (full), 'executive' (decision layer + thin pointers), or 'raw' (modules only).
Peer comparison (only on `peer-comparison` records)
comparison
Optional
Side-by-side benchmark against up to 3 peer domains across 8 metrics. Only present when input compareTo[] is populated — emitted as a separate dataset record with recordType: 'peer-comparison'.
Domain
domain
Optional
Company website domain
Company Name
companyName
Optional
Company researched
Research Date
researchDate
Optional
Date research was conducted (ISO YYYY-MM-DD)
Website
website
Optional
Website analysis: title, description, favicon, ogImage, socialLinks, security headers, robots.txt presence, sitemap URL, llms.txt presence, RSS feed URL
Wikipedia
wikipedia
Optional
Wikipedia summary and metadata
GitHub
github
Optional
GitHub presence: org profile, top repositories, total stars, language breakdown, npm/Docker org footprint
Financials
financials
Optional
SEC filings, ticker, CIK, exchange, address, SIC code
Research
research
Optional
Academic papers from OpenAlex (sorted by citation count)
Patents
patents
Optional
USPTO patent filings by assignee (count + recent grants)
DNS
dns
Optional
DNS records: A, MX, TXT, NS, CAA + parsed SPF/DMARC/DKIM status + email-provider classification
Subdomains
subdomains
Optional
Subdomain enumeration via Certificate Transparency logs (crt.sh)
Infrastructure
infrastructure
Optional
Wayback Machine first-seen date, security headers, .well-known/security.txt, ai.txt / llms.txt presence
Community
community
Optional
Hacker News mention count + top stories
Diff (since last run)
diff
Optional
Changes detected since previous run on same domain: new SEC filings, new GitHub repos, DNS changes, new TXT verification tokens, new subdomains. Only populated when scheduled / re-run on same domain.
Error
error
Optional
True when the run hit a fatal error before producing a report
Message
message
Optional
Human-readable error message (when error=true)
Timestamp
timestamp
Optional
ISO timestamp of error (when error=true)
Social Media
socialMedia
Optional
Social media presence across Twitter/X, LinkedIn, Facebook, Instagram, YouTube, GitHub