GigSoul AI Research Agent

Name the top 5 marketing trends for an industry or topic, with a short description of each. Use for trend briefings and idea generation. Based on model training knowledge, not a live lookup; figures are estimates. It may miss very recent shifts. For a full plan built on these, use generate-marketing

Build a full go-to-market plan for a product: target audience, positioning, channels, and timeline. Use when starting from scratch or needing a whole plan. For only a tagline and value proposition, use generate-positioning-statement. For a 30-day posting schedule, use generate-content-calendar.

Verify a paid deliverable against its offer, original request, payment receipt, and acceptance rules. Returns ACCEPT/REJECT/REVIEW with stable reason codes, per-check booleans, content hashes, and an evidence packet. JSON-only checks: schema, required fields, freshness, source presence, receipt bind

Disposable webhook test lab. capture: create a disposable HTTPS callback URL that records raw requests. retrieve: fetch captured requests. replay: send duplicate, out-of-order, delayed, malformed, or invalid-signature scenarios to your target URL. conformance: bounded webhook test battery with a mac

Validate and repair an n8n workflow JSON before production. static: deterministic plus model checks of nodes, connections, credentials, expressions, reachability. simulated: synthetic dry run with safe fixtures. repair: minimal RFC 6902 JSON Patch plus a replay fixture, re-validated in-worker. Retur

Map a market's competitors: names, strengths, weaknesses, and how each is positioned. Use for landscape and positioning questions ('who else does X and how do they differ'). Pricing is not covered; for plan tiers and price comparison use research-competitor-pricing. Based on model training knowledge

Suggest the top 10 SEO keywords for a topic with estimated relative search volume and difficulty. Use for early keyword ideation and content planning. Based on model training knowledge, not a live lookup; figures are estimates. Not live keyword-tool data; validate volumes in an SEO tool before commi

Write one LinkedIn post on a topic: post text, hashtags, and an opening hook. Use for a single original post. To adapt an existing article for LinkedIn and other channels at once, use repurpose-long-form. To score a draft before posting, use predict-viral-potential.

Write a first-touch cold email for a product and audience: subject line, preview text, and a 100-150 word body. Use for the opening message. For the follow-ups after no reply, use generate-follow-up-sequence. To check spam risk before sending, use check-email-deliverability.

Research a sales prospect company: company profile, decision-maker role profile (role-level, not a named person), likely pain points, and an outreach angle. Use before writing outreach; then use generate-cold-email. Based on model training knowledge, not a live lookup; figures are estimates.

Condense a document or article into what it says: main argument, key findings, and supporting evidence. Use when you need a faithful digest with no opinion added. For judgment on top of the content (insights, gaps, recommended actions), use analyze-document. To turn an article into social posts, use

Evaluate a document or article and say what to do about it: a short summary plus key insights, weaknesses or gaps, and concrete recommendations. Use when you need judgment or next steps. For a neutral digest only, use summarize-document (cheaper). For customer complaints inside a text, use extract-p

Turn one article into an X/Twitter thread of 5-10 posts. Use when X is the only target. For LinkedIn, email, and blog versions at once, use repurpose-long-form. URLs are not fetched; paste the text.

Repurpose one long piece for several channels at once: a LinkedIn post, an X thread, an email draft, and a blog intro. Use when you need all four. For an X thread only, use thread-from-article (cheaper). URLs are not fetched; paste the text.

Write podcast show notes from a transcript or outline: episode summary, key timestamps, guest bio, and links. Timestamps and links come only from the input; supply a transcript with times for accurate timestamps. URLs are not fetched; paste the text.

Pull the top 5 customer pain points out of text you supply (reviews, interview notes, support tickets, forum threads), each with severity and a suggested fix. Use when you have raw customer voice. For overall tone and emotion, use analyze-sentiment. For review trends over time, use analyze-review-ve

Classify the overall sentiment of a text and name its key themes and emotional triggers. Use for 'how do people feel about this'. For specific problems to fix, use extract-pain-points. For review momentum over time, use analyze-review-velocity.

Compare competitor pricing in a product category: competitor names, plan tiers and price points, and a value-for-money comparison. Use for pricing and packaging decisions. For strengths, weaknesses, and positioning, use analyze-competitors. Based on model training knowledge, not a live lookup; figur

Write positioning for one product: tagline, value proposition, and differentiation. Use for a single positioning block. For a full plan with channels and timeline, use generate-marketing-strategy. For competitor research first, use analyze-competitors.

Write a 3-email follow-up sequence (subjects, timing, and body) for prospects who did not reply to a first email. For the first-touch email, use generate-cold-email.

Write webinar invitation copy: two invite emails and a landing-page blurb. Use when promoting a scheduled webinar or live event.

Write 10 open-ended discovery questions, grouped by category, for interviewing a customer for a full written case study (problem, solution, results). For short quotable praise, use generate-testimonial-questions.

Write 8 short questions that draw out quotable customer testimonials. Use when you need quotes, not a long story. For a full case-study interview, use generate-case-study-questions.

Judge review momentum from review data you supply (dates, ratings, text): velocity score, sentiment trend, and response recommendations. Use for 'are reviews speeding up or slowing down, and how should we respond'. You must include the reviews; it does not fetch from G2, Capterra, or app stores. For

Plan a 30-day content calendar: daily themes, post types, and a posting schedule. Use for planning a month of content. For the overall strategy behind it, use generate-marketing-strategy. For the posts themselves, use generate-linkedin-post or repurpose-long-form.

Score a draft post before publishing: a virality score broken down by emotion, specificity, novelty, and shareability, plus the best-fit platform. A heuristic prediction, not a reach forecast. For landing-page copy use score-landing-page; to generate test variants use generate-ab-test-variants.

Create 3 A/B test variants of existing marketing copy, one variable changed per variant, each with a hypothesis, plus a test plan (primary metric, sample size, duration). Tone stays close to the source. Use when you already have copy to test. To score copy without variants, use score-landing-page or

Full deliverability audit of email copy: 0-100 score, every flagged phrase with its reason, subject-line risk, concrete rewrites, and SPF/DKIM/DMARC reminders. Use when a draft failed or warned in check-email-deliverability, or before an important send. For a quick pass/fail on many emails, use chec

Fast, cheap spam-risk gate for one email: pass/warn/fail, a 0-100 score, and up to 3 flagged phrases. Use for a go/no-go before sending or for bulk screening of many drafts. No rewrites. For a full audit with reasons, subject-line risk, and rewritten copy, use analyze-email-deliverability. Content s

Score landing-page copy for conversion: 0-100 with a 5-part breakdown (headline, value proposition, social proof, CTA, friction), the top 3 issues, and 3 quick-win rewrites. Use for a page's text. URLs are not fetched; paste the text. For social posts, use predict-viral-potential.