Koshima

It’s interesting how credits became the default abstraction for usage but without transparent burn rates, they're just tokens of confusion.

Couldn’t agree more. UPI is the unlock, not just the price. The credit card wall has quietly excluded 90% of the country even those curious and ready to pay.

Indeed it is!!

You’re right to be cautious. Perplexity’s throttling showed how fast “democratization” can become “degradation.”

But OpenAI might be running a playbook most competitors can’t afford to match: - Lock distribution now - Trade margins for data + ubiquity - Introduce upsells / infra plays later

100%. India has ~750M+ internet users and most of them have skipped the desktop era—straight to mobile + UPI.

With ₹399 and no credit card barrier, this could be AI's “Jio moment.” Suddenly, millions of first-time payers can access GPT-5.

Imagine the ripple effects on education, vernacular content, coding bootcamps, and small-town creators.

Absolutely. Perplexity with Airtel and Gemini with Reliance already hinted that India’s AI adoption will come through distribution moats.

Yes, you can downgrade from the $20 Plus plan to ChatGPT Go. The switch takes effect at the end of the current billing cycle.

OpenAI just launched a new plan called ChatGPT Go and exclusively for India.

₹399/month (roughly $4.80), with the support of UPI payments. And this wasn't a loud rollout at all, just a quiet addition to their pricing page. Now the point is that this isn't just a discounted GPT-Plus plan but it's actually built for India's mass market. Think of the students and everyday users living beyond tier 1 cities, who've never paid for AI before but know what ChatGPT is. The highlight is that along with credit cards they have added the support of UPI which is widely accepted across the country because let's face it that not everyone owns a credit card in India.

What’s included:

- GPT-5 (with extended usage)

- Image generation

- File uploads

- Python tools, memory, custom GPTs

What’s missing:

- GPT-4o or API access

- Connectors, Sora, or enterprise features

- No annual billing or bundles

They’re clearly not targeting the English-speaking dev crowd that already uses ChatGPT. This feels more like a test run for mass-market localization at scale, with India as the first sandbox.

$20/month doesn’t work in a country where Netflix costs ₹149. But ₹399 with UPI, is an unlock.

Feels like OpenAI is prepping for the next 500 million users, not the next 500 YC-backed teams.

Docs: https://help.openai.com/en/articles/11989085-what-is-chatgpt...

Would love to hear from anyone testing the usage caps.

Appreciate it. For token-based metering, we lean on event-level tracking with strict timestamps and unique identifiers to maintain billing precision, even when upstream responses are delayed or partial. If OpenAI or similar services provide incomplete data, we flag those events for retry or exclusion to avoid corrupting aggregates.

Retroactive adjustments are handled by versioning usage records, so instead of overwriting historical events, we apply corrections as delta events. Invoices pull from the latest state, but we preserve full audit logs underneath to avoid integrity gaps.

Would love to set that up. We’ve seen LLM infra teams hit billing issues fast because of usage spikes, token metering, credit handling. It gets messy to build it in-house.

I’ll DM you, we can run a quick POC and see if it fits your setup. Appreciate the interest.

I am working on building Flexprice(https://flexprice.io/), an open source monetization platform for AI and Agentic companies.

This week, we’re doing a 5-day launch week, where we’re shipping a new set of billing features every day. Github link: https://github.com/flexprice/flexprice

Appreciate that, and glad Flexprice showed up at the right time. We’ve seen the same pattern: credit systems always look simple with one plan, but as soon as you introduce volume tiers, expirations, or promo logic, it turns into fragile code fast.

We’re building Flexprice specifically to avoid that constant rewrite cycle, good to know it resonates. Happy to chat if you run into edge cases while scaling those new plans.

Good luck with your GenAI launch!

Appreciate it. We enforce idempotency at the event level using client-provided deduplication keys, so even with high concurrency or retries, the billing pipeline stays consistent.

For internal retries, we batch in-memory and attach unique IDs before dispatch to avoid double-counting.

Great question!

Our approach focuses on: - Fire-and-forget ingestion with in-memory queues so events don’t block product requests - Strict idempotency tokens tied to every event, enforced at the API layer - Lightweight retry logic that prevents double-counting but guarantees delivery under transient failures

Storage-wise, we’ve leaned on a mix of time-series DBs for raw events and pre-aggregated summaries for billing views.

Would love to swap notes on failure patterns or queue setups if you’ve dealt with similar scale.

One thing we’re handling differently is entitlements. Most billing tools stop at metering and invoicing, but they don’t track what features or limits a customer can actually access based on their plan, usage, or credits. We’re building that into the system so your app doesn’t need to maintain extra state for feature flags or usage limits.

I am building Flexprice, an open source metering and billing platform for AI and agentic-based companies. I've recently published a guide showing how to replicate Clay's credit-pricing model.

This guide includes: - Configuring recurring credit grants (e.g., 100 credits/month) - Capping rollover at 2× monthly allocation - Real-time metering (e.g., 10 credits to create a table; 1.5 per-row on enrichment) - Monthly vs annual billing models, credit expiry rules

This addresses a challenge many SaaS/AI/API products face: building transparent, usage-aligned pricing that’s easy to iterate on.

Would be grateful for HN feedback especially around edge cases or UI/UX when exposing credit consumption to users.

This is really cool! Building a browser engine from scratch is no small feat, especially when handling complex CSS features like calc(), var(), and percentage units. It’s a great way to learn the inner workings of the web.

Curious about your approach to the networking stack. Are you planning to support more protocols like HTTPS or WebSockets in the future, or is the focus more on keeping this lightweight and minimal for now?

not sure why but I think now school and colleges should promote students to use the new-age technologies. Using ChatGPT, they can shorten the time to research and do groundwork instead!

I’ve noticed this as well. It’s surprising how often the true value of certain high-quality items isn’t obvious until you’ve experienced them firsthand. It’s not just about status, but often about longevity, comfort, or simply a better user experience.

For example, things like handmade leather shoes, solid wood furniture, or even high-end kitchen tools like Miele or Sub-Zero appliances can feel like overkill until you’ve actually used them. Then you start to appreciate the craftsmanship, the reduced hassle, and the longevity they offer.

Curious if others have had similar experiences – what’s one “expensive” item that genuinely changed your perception once you owned it?

You’re definitely not the only one. I’ve seen a few discussions recently about unexpected billing for GitHub’s “code security” features. It seems like the feature gets auto-enabled without clear communication, and turning it off isn't straightforward.

One thing you can try is reaching out directly to GitHub support via their official Twitter account. Sometimes their social media team responds faster than traditional support. Also, check if any new team members or integrations could have triggered the feature.

It's fascinating how much body language and facial expressions differ across cultures. While some societies value open, expressive interactions, others lean toward reserved or neutral expressions. I’ve noticed this in my travels, a simple nod or slight smile can mean very different things depending on where you are.

I wonder if these cultural norms around eye contact and facial expressions have roots in deeper societal structures, like the emphasis on individualism vs. collectivism, or even the pace of life in different regions.

What do you think? Could these small, often overlooked gestures reflect much larger cultural attitudes?

It’s a great question. I think part of the answer lies in changing how we frame entrepreneurship. Instead of just focusing on the hustle and financial rewards, we should highlight the creative problem-solving, the freedom to experiment, and the satisfaction of building something meaningful from scratch.

A lot of the excitement in other subjects comes from discovery and exploration, and starting a business can be just as much about learning and adapting as it is about scaling and profits.

What do you think? Would reframing entrepreneurship as a craft make it more interesting to beginners?

Are you building metering and billing in-house?

What challenges are you currently facing in monetizing your platform?

Ever happened that the growth or product manager comes in between the sprint and tell you that you've to make changes in the current pricing plan?

How are you managing it?

It’s a reminder that financial inclusion isn’t just about tech – it’s about creating systems that can handle the messy reality of real-world transactions. It’s disappointing that a platform like BuyMeACoffee, which is supposed to empower creators globally, is now cutting off entire regions.

Honestly, the best way to make your iPhone worse is to just turn on every notification and location permission. You'll go from "always connected" to "always interrupted" in no time.

i really need this!

I’m not sure if I trust any billionaire to solve loneliness, but at least he is building platforms that keep people connected...

I think it’s fair to push for clear attribution in these cases, but it’s also important to remember that the MIT license is intentionally permissive. It was designed to make sharing code easy without too many hoops. If Ollama is genuinely trying to be part of the open-source community, a little transparency and acknowledgment can avoid a lot of bad blood.

It's fascinating how quickly the ecosystem around LLM agents is evolving. I think a big part of this "unreasonable effectiveness" comes from the fact that most of these tools are essentially chaining high-confidence steps together without requiring perfect outputs at each stage. The trick is finding the right balance between autonomy and supervision. I wonder if we'll soon see an "agent stack" emerge, similar to the full-stack frameworks in web development, where different layers handle prompts, memory, tool calls, and state management.

This is a classic case of IP abuse, and it's tough to ignore. If the company has been using your work without a license for a decade, that’s a huge liability on their side. It might be time to remind them that open source is not free labor, and they can’t just brush off 10 years of unpaid work. At the very least, they should come to the table for a serious negotiation.

The timing makes sense if you consider the broader trend in the LLM space. We're moving from just text to more integrated, multimodal experiences, and having a tightly controlled engine like this could be a game changer for developers building apps that require real-time, context-rich understanding.