MINDSCAST
Mindscast is your AI content-marketing employee that actually knows your product. It first builds a Product Context Document (PCD) — a working memory of your features, proof, pricing, customers, pains, competitors, and forbidden claims — then turns that memory into LinkedIn and X content you can approve.
3 ingest paths · ~30-second grounded extraction · 0 hallucinated features internal benchmarks, July 2026
Extracting your PCD is free — passwordless sign-in, no card required.
01 REVEAL
The PCD is the product memory a good marketer would build before writing: what the product does, who feels the pain, what proof exists, who you compete with, and what claims are off-limits. The graph matters because content should start from a real customer problem and trace back to approved product facts.
SELECTED CHAIN
Mindscast → feature → buyer pain → proof
Click any node to see the narrative chain Mindscast would use to explain its own product.
The same content the interactive graph shows, in plain text: every node of the sample Product Context Document, and the narrative chain Mindscast would use to explain it.
Mindscast product
Mindscast PCD: one product memory, many narrative paths
This simplified PCD shows how Mindscast explains itself: features sit together, audiences and pains sit together, proof and guardrails stay connected, and the dashboard turns the graph into a daily workflow.
Product → features → audiences/pains → proof + guardrails
PCD builder feature
PCD builder → generic-post pain → founder-led team
The PCD builder extracts features, pricing, proof, audience pains, differentiators, and forbidden claims before writing. That is why Mindscast does not start from a blank prompt.
PCD builder → solves generic AI posts → founder-led team
Asset parsing feature
Asset parsing → asset chaos → reusable content inputs
Mindscast should understand more than text. Multimodal asset parsing turns screenshots, creative assets, product images, and source files into recognized product facts and usable campaign material.
Asset parsing → solves asset chaos → feeds grounded campaigns
Grounded posts feature
Grounded posts → trust pain → proof
Posts are generated from approved PCD facts. The product claim, buyer pain, and proof point stay connected, so users can see why a post says what it says.
Grounded posts → solve untrusted claims → cite 0 fake features
Dashboard feature
Dashboard → daily publishing workflow → less manual prompting
The dashboard makes the PCD operational: review grounded drafts, schedule LinkedIn/X content, inspect proof, and keep campaigns moving without rebuilding context every time.
Dashboard → review/schedule → consistent founder-led publishing
Founder-led team audience
Founder-led team → pains → Mindscast features
The target user is a small team or solo founder that needs consistent product marketing without hiring an agency or prompting a generic model every day.
Founder-led team → generic posts + untrusted claims → PCD + grounded posts
Solo marketer audience
Solo marketer → asset chaos → multimodal product memory
A small marketing team has screenshots, landing pages, pitch decks, and product visuals scattered everywhere. Mindscast should recognize those assets and make them reusable in campaigns.
Solo marketer → asset chaos → asset parsing
Generic posts pain
Pain: generic AI posts
Generic posts happen when the model does not know the product. Mindscast fixes that by extracting the PCD first, then writing from approved product memory.
Generic AI posts → solved by PCD builder → Mindscast
Asset chaos pain
Pain: asset chaos
Product teams already have useful visuals, screenshots, launch assets, and proof artifacts. Multimodal parsing helps Mindscast identify what each asset shows and where it belongs in the product story.
Asset chaos → asset parsing → recognized campaign material
Untrusted claims pain
Pain: untrusted claims
A founder cannot publish claims that invent proof, pricing, or capabilities. Mindscast writes from grounded PCD chains and keeps unsupported claims out.
Untrusted claims → grounded posts → proof + guardrails
0 fake features proof
Proof point: 0 fake feature ids
This proof makes the product promise concrete: Mindscast can reject generated content that references features not present in the PCD, and asset-derived claims still need evidence.
0 fake features → supports grounded posts + asset parsing
Product memory differentiator
Differentiator: product memory beats prompt-only AI
Mindscast's differentiator is not just better wording. It keeps a structured product memory that connects features, assets, pains, proof, competitors, and forbidden claims.
Product memory → compared against prompt-only AI
Prompt-only AI competitor
Alternative: prompt-only AI
Prompt-only AI can write fluent copy, but it does not maintain product memory, source links, buyer pain chains, or forbidden-claim guardrails by default.
Prompt-only AI → compared with grounded posts
No invented proof guardrail
Guardrail: no invented proof
Forbidden claims are part of the PCD. They stop Mindscast from using impressive but unapproved proof, numbers, integrations, or outcomes.
No invented proof → protects product trust
A PCD is not a database for engineers. It is the product memory your AI marketer uses to choose an angle, cite proof, and stay inside approved claims.
Paste your homepage. We crawl pricing, features, about, and docs — ~30 seconds to a full graph.
Install the GitHub App and pick the files that describe your product. Read-only, repo-scoped.
Your own Claude Code / Cursor / Codex reads the repo locally and uploads only the structured PCD you approve. Source code never leaves your machine.
02 POWER
Product outcomes from our July 2026 extraction and grounding benchmarks.
| Spec | Measured | Source |
|---|---|---|
| Time to product memory | 25–42 s | 4-site homepage benchmark |
| Relevant feature selection | 100% | marketing-filter benchmark; 3.3× over raw LLM |
| Invented feature references | 0 | post-generation hard-filter validation |
| Evidence relationship precision | 82% | hand-audited, 5-provider benchmark |
| Source-page price capture | Verbatim | tested across subscription, usage-based, and transaction pricing |
When a source page includes a product price, Mindscast preserves the published amount and wording instead of rounding or rewriting it. We tested this across subscription, usage-based, and transaction pricing in our cross-industry extraction trial.
03 LAUNCH LINEUP
Mindscast turns each graph chain into a marketable angle: pain first, product fact second, proof when available, and a CTA tied to why that customer buys. You approve the product memory once; every post stays inside it.
ANGLE: fear of surprise cloud bills after shipping fast
Shipping AI code faster is great until a hidden cryptominer turns Saturday into a cloud-bill incident. Mindscast would make the product story concrete: kernel-level eBPF detection catches the behavior code review missed, and the rollback fix is one click away.
From PCD chain: eBPF detection → cryptominer cloud bill → funded startup team
X
ANGLE: founders need investor language without consultant pricing
A founder does not need another spreadsheet template. They need the valuation story an investor can understand. The post can lead with the pain, then cite the real product fact: 9 valuation models in one report, from pre-seed to Series B, priced at €30 instead of a €25k advisory bill.
From PCD chain: 9 valuation models → valuation cost too high → early-stage founder
ANGLE: compliance deadline as a reason to start practical AI training
A compliance post does not need to sound like a legal memo. It can start where the buyer is: Article 4 is already in force, the team has not started, and the offer is training that satisfies the obligation while teaching people what AI can actually do.
From PCD chain: AI literacy training → Article 4 obligation → compliance-conscious SME
04 EDITIONS
A part-time content marketer runs $1,000+/month. A Mindscast employee that actually knows your product starts at €29.
SOLO
€29/month
Pay monthly after your 14-day trial.
14 days of the complete Solo plan
€0 today. Card required; then €29/month after the trial.
Cancel before the trial ends and pay nothing · checkout secured by Stripe
What's free, forever: extracting and editing your PCD — from a URL, a GitHub repo, or your local coding agent via MCP. The trial gates content generation, not understanding your product.
05 FAQ
Every post is generated from a Product Context Document (PCD): the product memory Mindscast builds from your own sources. It contains approved features, buyer pains, proof points, pricing, competitors, and forbidden claims. A post-generation hard filter drops any feature id not present in that memory. In our benchmark the marketing filter reached 100% on-topic feature selection, 3.3× better than handing a raw LLM the full context.
Yes. The MCP ingest path runs inside your own Claude Code, Cursor, or Codex session: your local LLM reads the repo on your machine and uploads only the structured PCD JSON you approve. Source code never leaves your laptop.
A PCD is the working product memory a good marketer would build before writing: 15–30 features with proof points, audience segments with their pain points, competitor differentiators, verbatim pricing, and the claims you never want marketing to make. Posts trace claims back to approved facts in that memory.
URL crawl (paste your homepage, ~30 seconds), GitHub App (pick files from a public or private repo), or MCP (your local coding agent extracts from the full repo — highest fidelity, code never leaves your machine).
Solo is €29/month or €290/year. Both include up to 60 generated LinkedIn and X drafts per calendar month and multiple product PCDs. The same monthly allowance applies during the 14-day trial and throughout the paid subscription; it is not a trial-only allowance. A draft generated for both LinkedIn and X counts as two posts.
Extracting and editing your PCD is free and needs no card. The 14-day trial — which unlocks content generation and publishing — collects a card at checkout via Stripe; cancel anytime during the trial and you pay nothing. We say this up front because surprise charges are exactly the kind of claim we'd put on a forbidden list.
LinkedIn and X (Twitter) natively, with human-in-the-loop review gates before anything is published.
A product homepage ingested in 25–42 seconds in our July 2026 four-site benchmark. The result is an editable Product Context Document containing the product facts and relationships Mindscast uses for content generation; extraction time varies with site size and crawlability.
Yes — the URL path crawls your homepage plus pricing, features, about, and docs pages. In a July 2026 trial across tally.so, elevenlabs.io, posthog.com, and lemonsqueezy.com, every crawlable site passed the full extraction quality baseline, including verbatim capture of subscription, usage-based, and transaction pricing.