Postessia Selected for the Claude Startup Program and the Sarvam Startup Program

We have two pieces of news, and we wanted to share them together because they belong together: Postessia has been selected for the Claude startup program, and Postessia has been selected for the Sarvam startup program.

For a small team building an AI ghostwriter for LinkedIn out of India, being selected for the Claude startup program and the Sarvam startup program in the same season is a real milestone. One program comes from Anthropic, the company behind Claude, one of the strongest families of language models in the world. The other comes from Sarvam AI, an Indian company building foundation models with Indian languages and Indian users at the centre. Having both look at what we are building and decide to back it tells us that the problem we care about, LinkedIn posts that are actually worth reading, is a problem worth solving properly.

This post is longer than our usual announcement. We want to explain what each program is, why Postessia applied, what our goal really is, why we think most AI-written LinkedIn content is broken, what we are doing about it, and what all of this means for you if you use Postessia. And we want to say thank you, properly, to the people who got us here.

TL;DR

• Postessia, an AI ghostwriter that writes LinkedIn posts in your own voice, has been selected for the Claude startup program run by Anthropic.
• Postessia has also been selected for the Sarvam startup program run by Sarvam AI, an Indian foundation model company.
• Our goal is simple to state and hard to do: make LinkedIn post generation produce posts that are better to read, not just faster to make.
• LinkedIn is full of AI slop. Copying a prompt into a chatbot and pasting the answer into LinkedIn is not ghostwriting, it is just a bad prompt. We are putting real research into what makes a post sound like a specific person and read well.
• Nothing changes about your data, your voice profile or our pricing because of these announcements.
• If you want LinkedIn posts that sound like you, you can try Postessia free today.

What is the Claude startup program?

Claude is the family of AI models built by Anthropic. Alongside the models, Anthropic runs a startup program for early-stage companies that build their products on Claude. Startups apply, and the ones selected get support to build and grow on Claude.

Why does a program like this exist? Because building a reliable AI product is nothing like calling an API once and shipping. Every change to a prompt, every evaluation run, every new feature and every failed experiment costs real model usage long before it reaches a paying customer. For a young company, support at that stage is the difference between testing an idea properly and shipping the first thing that roughly works.

Being selected for the Claude startup program is not a guarantee of anything by itself. It is a starting line. But it does mean that the team behind one of the best reasoning models available looked at a LinkedIn AI ghostwriter from India and decided it was worth backing.

What is the Sarvam startup program?

Sarvam AI is an Indian AI company building foundation models, speech and language technology built for India, with a strong focus on Indian languages and the way Indian users actually speak, write and work. Sarvam runs a startup program to support Indian startups building AI products on its platform and in its ecosystem.

For Postessia, being selected for the Sarvam startup program matters for a reason that goes beyond support. Postessia started with built-in context for Indian creators, founders and agency owners, because that is who we know best. Indian professionals write LinkedIn posts in a particular way: a mix of English and Hinglish, references that only make sense if you have worked in Bengaluru or Gurugram, a tone that is direct without being American-style loud. Most AI writing tools were trained and tuned on a very different internet. Being close to a company that is building models for India gives us a better path to serving Indian writers well.

Why both programs matter for an AI ghostwriter

People sometimes ask why we would join two programs instead of picking one. The honest answer is that they help with different parts of the problem.

• The Claude startup program helps us with the hardest reasoning work in Postessia: planning an angle, auditing a draft against a person's real voice, and catching the subtle generic patterns a human editor would strike out. That work benefits from strong models and from room to run many experiments.
• The Sarvam startup program helps us stay close to Indian language, Indian context and the Indian startup ecosystem, which is where most of our early users are and where the next wave of LinkedIn creators is coming from.

An AI ghostwriter has to be good at two things at once: understanding people, and writing like a specific one of them. Being selected for the Claude startup program and the Sarvam startup program supports both.

The problem we are actually trying to solve

Here is the thing we keep coming back to.

There is so much slop on LinkedIn right now.

Open your feed and you will see it within three posts. The "I am humbled to announce" opener. The one-line paragraphs with a dramatic pause after every sentence. The "Here is what I learned" list that could have been written by anyone about anything. The question at the end that nobody wants to answer. LinkedIn readers have learned to recognise these posts in seconds, and LinkedIn itself is now testing a way for readers to report content that seems like AI slop.

Most of that slop comes from the same workflow. Someone opens a general-purpose chatbot, types "write a LinkedIn post about leadership lessons from my startup", copies the answer, and pastes it into LinkedIn. Maybe they change a word or two. Then they wonder why the post gets no engagement and why it reads nothing like them.

We want to be very clear about this, because it is the reason Postessia exists:

Copy-pasting from a chatbot is not AI ghostwriting. It is just a bad prompt.

A general-purpose model given a one-line request will produce the statistical average of every LinkedIn post it has ever seen. That average is exactly what slop looks like. It is not the model's fault. It is doing what it was asked. The problem is that nobody told it who is writing, how they write, what they actually believe, what happened in the story, or what a reader is supposed to take away. Without that, the output is generic by design.

A real ghostwriter does not work that way. A good human ghostwriter spends the first few weeks reading everything a client has written, listening to how they talk, learning the phrases they use and the ones they would never say. Only then do they write a single line. That is the standard Postessia holds itself to, and it is why we describe Postessia as an AI ghostwriter rather than a post generator.

Our goal: posts that are better to read

Our goal is not to help people publish more. LinkedIn does not need more posts. Our goal is to make LinkedIn post generation produce posts that are better to read: posts that sound like a real person, say something specific, and respect the reader's time.

"Better to read" is a measurable idea for us, not a slogan. We look at things like:

• Does it sound like the writer? If you put the draft next to the person's last ten posts, can you tell which one the AI wrote? The answer should be no.
• Does it say anything? A post that could be published by a thousand other people is not a post, it is noise. Every draft needs a specific claim, a specific moment or a specific opinion.
• Is it honest? No invented statistics. No made-up client stories. No "studies show" without a study.
• Does it read well on LinkedIn? Where does the "see more" fold land? Does the hook earn the click? Is the structure something a person would actually write, or a template?

Getting these right takes research, and that is where we have been spending most of our time.

What "putting research on it" actually means

It is easy to say you are doing research. Here is what it looks like for us in practice.

1. Studying how real people write before writing anything

Before Postessia writes a single word for you, it studies your existing posts. Not just the topics, but the mechanics: how long your sentences run, how you open, whether you use questions or statements, which words you reach for repeatedly and which ones you never use. That becomes a voice profile.

Most of our research time goes into making this profile sharper. We compare drafts against the writer's real posts, look for drift toward generic phrasing, and tune until the difference is hard to spot. The goal is not a "professional LinkedIn tone". The goal is a draft your colleagues would believe you wrote yourself.

2. A three-step pipeline instead of one prompt

Every Postessia post goes through a quality pipeline rather than a single shot at a prompt:

• Planner shapes the angle and decides what the post is actually about, so the draft has a point of view instead of a topic.
• Drafter writes the post in your voice, using the profile above.
• Auditor reviews the draft and strips out fluff, filler and anything that does not sound like you.

Each step is a separate job with its own instructions and its own checks. This is the kind of work where strong reasoning models earn their keep, because each stage has to follow instructions precisely and, just as importantly, know when to leave a sentence alone. Being selected for the Claude startup program gives us more room to run these stages carefully instead of cutting corners.

3. An AI gatekeeper that blocks slop before you see it

The last pass in Postessia is what we call the AI gatekeeper. It blocks the patterns that make AI-written LinkedIn posts easy to spot: clichés, vague claims, invented numbers, hooks that promise more than the post delivers, and phrasing that is off-voice for you specifically. If a sentence could appear in a thousand other posts, it gets flagged.

This matters more every month. As LinkedIn gets better at recognising low-effort AI content, the cost of publishing it goes up. A tool that protects your reputation is worth more than a tool that writes fast.

4. Real Story Mode, with a strict no-invented-facts rule

Some of the best LinkedIn posts begin as a messy note: a client call that went sideways, a hiring mistake, a launch that taught you something. Real Story Mode takes that raw material, whether it is bullet points, a voice-note transcript or a few lines from your notes app, and turns it into a structured post with a clear hook and takeaway.

The rule we hold ourselves to is simple: Postessia should never invent facts. If you did not say it happened, it does not appear in the post. Keeping a model faithful to the source material is one of the hardest problems in AI writing, and it is one of the areas where we plan to keep investing research time.

5. A LinkedIn preview that respects the fold

Reading well on LinkedIn is not only about the words. It is about where the "see more" line lands, how the first two lines look in a crowded feed, and whether the formatting feels human. Postessia shows you the post exactly as it will appear on LinkedIn before you publish, so you can fix the hook, not just the body.

Why an Indian AI ghostwriter needed both programs

We are building Postessia from India, for Indian creators first and for writers everywhere after that. That shapes the research in ways that are easy to miss if you only look at the product from outside.

Indian LinkedIn has its own rhythm. Founders write about fundraising in a way that references Indian investors and Indian markets. Consultants write in a register that sits between formal and conversational. Many creators switch between English and Hinglish inside a single post, and they do it on purpose. A tool tuned entirely on an American internet flattens all of that into one voice.

Being selected for the Sarvam startup program puts us close to a company that cares about exactly this: how Indians actually speak and write. Being selected for the Claude startup program gives us access to the reasoning quality we need for the planning and auditing steps. Together, they let a small team build an AI ghostwriter that is both deeply capable and genuinely local, and then make it work for creators anywhere, which is why we offer pricing in both INR and USD.

What changes for Postessia users

Short answer: the product gets better, and nothing you rely on gets worse.

• Better drafts over time. More room to experiment means more evaluation of voice matching, the quality pipeline and the AI gatekeeper, and faster improvements to each.
• No change to your data. Your posts, your voice profile and your drafts are handled exactly as described in our privacy policy. Being in either program does not change who can see your content.
• No change to pricing because of this. Plans and prices are listed on our pricing page, and you can still start on the free plan with no card required.
• Same focus. We are still building one thing: LinkedIn content in your own voice, for creators, founders and professionals building a personal brand.

Who Postessia is built for

Postessia is an AI ghostwriter for LinkedIn, and it is designed for people who care how they sound:

• Founders who know LinkedIn matters for hiring, fundraising and sales, but cannot spend five hours a week writing.
• Creators building a personal brand who want consistency without losing their voice.
• Consultants and freelancers who win work through trust, and know that a generic post costs more trust than it earns.
• Agency owners and ghostwriters who are tired of tools that make every client sound like the same template.

If you have ever deleted a draft because it "sounded like AI", Postessia was built for you.

What we are building next

Being selected for the Claude startup program and the Sarvam startup program does not change our roadmap so much as speed it up. The areas we are focused on next:

• Sharper voice profiles that capture tone differences between, say, your thought-leadership posts and your behind-the-scenes posts.
• Stronger quality checks, so the Auditor and the AI gatekeeper catch more subtle generic patterns, not just the obvious ones.
• Better Real Story Mode, so rough notes become strong posts with less editing, while staying strictly faithful to what actually happened.
• More useful topic variations, so one idea becomes several genuinely different angles instead of three versions of the same post.
• Better Indian context, so Hinglish, Indian references and the Indian professional register are handled as a feature, not an edge case.

We will share progress on each of these here on the blog as they ship.

A thank you

We want to end with the part that matters most to us.

To everyone who signed up early, sent feedback, reported a draft that sounded off, argued with us about a hook, or told a friend about Postessia: thank you. Most of the improvements in the product today came from those conversations. Being selected for the Claude startup program and the Sarvam startup program is a milestone for the team, but it belongs to you as much as to us. Programs select products, and the product exists because you used it and told us the truth about it.

To Anthropic, thank you for backing a small team with a specific idea about how AI writing should work. The research we described above is only possible because we can afford to run it carefully.

To Sarvam AI, thank you for building for India and for backing Indian founders who are doing the same. It means a great deal to be selected by a company that understands why "sounds like you" means something different in Mumbai than it does in San Francisco.

And to the LinkedIn writers who refuse to publish slop, even when it would be easier: thank you for setting the bar. You are the reason we think an AI ghostwriter should make people sound more like themselves, not less.

If you have not tried Postessia yet, you can start for free and generate your first posts in your own voice in a few minutes.

Frequently asked questions