Pangram, a 24-person AI-detection startup headquartered above a Popeyes in Brooklyn, has raised $13M to date and is now the tool cited when book deals collapse, prize winners get challenged, and Substack posts get scanned. In July 2026 the company raised $9M and shipped Pangram 4, and it has posted six new job openings that would grow headcount by 25%. Its total funding is roughly 0.0072% of what OpenAI has raised, yet its scores are moving decisions inside publishers that command far larger balance sheets.
The product outputs a single percentage — a best-guess estimate of how much of a given text was AI-generated. In January, CEO Max Spero posted on X that Mia Ballard's self-published novel Shy Girl scored 78% AI-generated; Hachette had picked the book up for traditional release and then canceled it. Pangram has since scored a New York Times Modern Love column at 100%, a Commonwealth Short Story Prize winner at 100%, the novel Daggermouth at 60%, and a thriller titled Call Me, I'll Hide the Body — which had sold for $2.4M — at 97%.
In late July, Substack integrated Pangram into its platform so readers can check any post. A Substack representative said the company's philosophy is not anti-AI, but that readers should know what they're consuming. Substack declined to disclose whether it pays Pangram per scan.
Key facts
- 01Pangram raised $9M in July 2026, bringing total funding to $13M across a 24-person team based above a Popeyes in Brooklyn.
- 02Pangram scored Mia Ballard's novel Shy Girl at 78% AI-generated; Hachette canceled the release shortly after.
- 03Substack integrated Pangram in late July 2026 so readers can scan posts for AI-generated text.
- 04A Stony Brook study using Pangram found nearly 20% of 14,419 self-published novels carried substantial AI-detection scores.
- 05Pangram flagged a Modern Love column at 100%, a Commonwealth Short Story Prize winner at 100%, and a $2.4M thriller at 97%.
Spero, 30, cofounded the company with Bradley Emi after the two met at Stanford. Spero worked on Google's FLoC cohort-targeting effort before it was killed in 2022, then joined the autonomous-vehicle company Nuro; Emi did stints at Tesla and the AI biotech firm Absci. They launched Checkfor.ai in 2023 and renamed it Pangram a year later, entering a crowded detection market that already included Originality.ai, GPTZero, and Turnitin. Pangram outperformed rivals in early independent testing and pulled ahead.
“All of our datasets are properly licensed, which I think is kind of rare in the AI world today.”— Max Spero, Pangram cofounder and CEO
The technical approach is what Pangram calls synthetic mirroring: take human writing, have an LLM generate a close match, and train the detector on the pair. The company also runs hard-negative mining, searching datasets for false positives it can mirror and feed back into training. Pangram's model is far less data-hungry than a frontier LLM, which is part of why Spero can claim the training data is fully licensed.
The Shy Girl case put Pangram on the map, but the mechanics were messier than Spero's public account. A Pangram account executive discussed the manuscript with a publishing analyst, who brought it to The New York Times. Critics, including The Drey Dossier investigative project, noted that Spero's copy of the manuscript came from a pirating site. Asked about that, Spero said he hadn't read the whole PDF and just fed it straight into Pangram.
The company has kept flagging high-profile targets. After the Commonwealth Short Story Prize winner returned a 100% score, Pangram ran every winner since 2012 through its model and identified three more it deemed suspect. Spero downplays the tool's influence on canceled book deals, arguing the score is only a small part of a broader publisher decision.
The market case is clear enough: AI writing is spreading through publishing faster than the industry can absorb. A Gotham Ghostwriters poll of 1,481 working writers found 61% use AI tools and 7% have published AI-generated text. Tuhin Chakrabarty, a Stony Brook computer science professor, used Pangram to analyze 14,419 self-published novels and found nearly 20% carried substantial AI-detection scores — the same dataset that surfaced the Daggermouth flag.
Big Five publishers are cagey about their use of the tool. Penguin Random House confirmed its editors may use approved AI-detection tools as one additional signal, but said they are not determinative. Simon & Schuster and HarperCollins declined to comment; Hachette and Macmillan did not respond. Some literary agents are running detection privately, quietly killing deals that never surface publicly.
The counterweight is a real credibility problem. Jane Friedman, an author and publishing expert, said there is distaste and anger toward detection software, with writers feeling the vendors are as harmful as the AI companies themselves. Chakrabarty, whom Pangram supplies with API credits and who describes Spero as a close friend, cautions that Pangram should not be the de facto judgment even as he defends it publicly. A single percentage — produced by an AI trained on synthetic mirrors of human writing — is now capable of ending a career before an editor finishes reading a manuscript.
Pangram sits in an uncomfortable spot: a detection company whose method depends on the same generative models it exists to police, whose most cited data point traces back to a friend's academic paper, and whose commercial future depends on publishers, universities, and platforms treating its output as reliable. The Substack deal is the first at-scale distribution test, and the $9M round funds the sales motion to sign more platforms on the same terms. If Pangram becomes the default filter for creative writing across Substack, education, and the Big Five's editorial workflows, its 78% and 97% scores will function less like evidence and more like verdicts — which is exactly the outcome its critics fear and its investors are betting on.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
