Shipping Day! We’re all proud and excited @nutrientdocs to officially release our PDF to Markdown converter. On our 200 document hand annotated markdown based benchmark it beats out every other major parser for speed and reading order while matching accuracy. We are shipping it as a closed source, free to use for up to 1,000 documents per month, no license key required, and your documents stay completely local. Read the full post
Nutrient
Clickstrike ran X creator campaigns around three Nutrient developer launches: a PDF-to-Markdown CLI, open-weight grounding models, and Document Authoring AI, reaching over 900K views.
900K+
Views across three launches
100K+
Views on each of the top creator posts
3
Developer launches amplified
The launches
Every launch, replayed.
Three Nutrient releases, each with the creator wave we ran around it. The Document Authoring AI creators posted their own demos rather than quoting a Nutrient post.
Original post
Creator wave
June 2026Most PDF parsers destroy formatting. Here’s how I fixed that for my AI workflows 🧵
PDF parsing almost killed my RAG pipeline 💀 I tested @nutrientdocs's PDF-to-Markdown CLI on a 40-page research paper, tables, inline citations, and multi-column layouts. One command. Clean, structured Markdown. Zero cleanup. Here's why that matters 🧵
The whole wave
- Creator post views
- 400K+
- Views on the top post
- 100K+
Original post
Everyone keep talking about how LLMs are good enough now. The LLMs have always sounded confident. I don't buy it. After years running and building document workflows @nutrientdocs (formerly @PSPDFKit), the breakage is in the numbers. A model confidently hallucinates answers from a filing table it never fully saw because a 512-token window chopped the table in half. So we open-sourced grounding-en, a 0.4B model on @huggingface that scores whether a claim is actually entailed by the source. On our open benchmark our two models sit at #1 and #2. Read the full post
Creator wave
Aug to Sep 2026MOST DOCUMENT AI GIVES YOU AN ANSWER. NUTRIENT GIVES YOU PROOF LLMs sound confident whether they’re right or wrong. in claims, mortgages, legal forms, a confidently wrong number is worse than no answer at all → every extracted value comes with a bounding box, confidence score and page index → hover a field, watch the highlight land on the exact pixels it came from → .92 AUC on number grounding vs .45-.66 for standard NLI models → open-sourced the grounding model, published the benchmark, they sit #1 and #2 on their own public leaderboard a highlight in the wrong place is a visible audit trail. that’s the whole point
I tested @nutrientdocs on a CMS-1500 health claim, which is exactly the kind of document that can trip up extraction systems. The red grid alone can get mangled by scanners, and the form is packed with dense fields where one wrong value can matter. What stood out to me was the grounding. Hover over any extracted field and Nutrient shows the exact place on the original claim where that value came from. Read the full post
The whole wave
- Creator post views
- 240K+
- Views on the top post
- 100K+
Creator wave
Aug to Sep 2026So I let an AI edit a real NDA, and it left a comment on every single change. @nutrientdocs Document Authoring AI opens the NDA in your browser and reviews it clause by clause. Each suggestion becomes a tracked change, and the comment next to it says exactly why the AI made that call. Read the full post
WHAT USED TO BE A JUNIOR ASSOCIATE'S JOB AT 2AM NOW TAKES SECONDS not saying it replaces the legal team I mean the first pass: filling blanks, checking standard clauses, getting the draft ready for the lawyer to review every edit lands as a tracked change with a comment explaining why using @nutrientdocs Document Authoring AI, the junior is still there. just reviewing instead of filling in blanks
The whole wave
- Creator post views
- 290K+
- Views on the top post
- 80K+
Overview
Nutrient, formerly PSPDFKit, builds document SDKs and APIs for developers who view, edit, extract from, and generate documents inside their own products.
In 2026 the company shipped a run of AI-era developer products: a PDF-to-Markdown converter for LLM and RAG pipelines, open-weight grounding models on Hugging Face that check whether an extracted claim is supported by the source, and Document Authoring AI, which brings reviewable AI editing into the document editor.
Clickstrike ran X creator campaigns around three of those releases between June and September 2026, alongside a press release placement on CIO Dive for the Data Extraction API launch.
The Challenge
Nutrient sells to developers and technical teams who build document workflows. That audience wants to see a tool work on a real document before it takes a claim seriously.
Each release also needed its own angle. A free CLI for developers, an open-weight model for ML engineers, and an AI editing feature for legal and document-heavy teams would not land with one generic message.
The releases also came close together. The grounding models and Document Authoring AI campaigns started on the same day, so the same creators had to carry two different stories at once.
The Solution
We built a bench of AI and developer creators on X and briefed them release by release, working from core messaging set with Nutrient at the start of the engagement.
Creators used the products before they posted. The CLI posts show real PDFs going in and clean Markdown coming out. The grounding posts test extraction on documents like a CMS-1500 health claim. The Document Authoring AI posts walk through an AI review of a real NDA, with every edit landing as a tracked change.
Every post was in the creator's own voice and format: single posts, threads, and quote posts of Jonathan Rhyne's own launch announcements. Two of the creators post in Spanish, which took the launches into Spanish-language AI audiences as well.
The bench carried over. Most of the creators from the CLI campaign came back for the grounding models and Document Authoring AI.
The Results
Across the three campaigns, creator posts reached over 900K views on X, and each launch reached developers through creators who had used the product first-hand.
Standouts included @zaynmcps's post on grounded extraction and his PDF-to-Markdown how-to thread, which each passed 100K views, and @Rixhabh__'s NDA walkthrough for Document Authoring AI, at more than 80K.
Earned media
The Data Extraction API launch ran as a press release on CIO Dive, which covers enterprise IT strategy and technology for CIOs.
900K+
Views across three launches
100K+
Views on each of the top creator posts
3
Developer launches amplified
Get Results Like Nutrient
Ready to grow your AI company? Let us build a custom marketing strategy for you.
Results
More Case Studies

Anam AI
We amplified Anam AI's Cara-4 launch video on X and turned it into a trending moment: over half a million views, with no paid advertising behind it.

Patronus AI
Clickstrike amplifies Patronus AI's announcements on X. The out of stealth launch reached 246,844 views across two announcement posts and 12 hand-picked creators, and Patronus AI has since brought Clickstrike back for a second campaign that hit number one in X's Today's News, plus three more campaigns now underway.

Neverbell, by Beam
How Clickstrike launched Neverbell from a brand new X account to 688,727 campaign views and 50 earned media placements, all anchored to a single coordinated launch day.
Ready to Be Our Next Success Story?
Let us create a custom marketing strategy for your AI company.