Hey folks, hope everyone's doing well. I've been testing out Searchable for a client project where we're migrating about half a million PDFs and support articles into a searchable knowledge base. The 14-day trial has been fantastic for setup and initial indexing – the API is clean and the dashboard is intuitive.
Now we're coming up on the end of the trial, and I'm trying to nail down the ongoing costs for a production index of this size. Their pricing page talks a lot about "records" but I know from experience that storage and operations can vary based on actual document size and query volume.
Could anyone who's running a similar-sized index with Searchable share what your actual monthly bill looks like? Specifically:
* What tier or plan are you on?
* Is the 500k count based on actual documents, or is there a record multiplier for text chunks?
* Any major cost drivers besides the pure document count? (e.g., search query volume, storage overages)
* How's the performance at this scale for those of you using it daily?
Just trying to budget accurately before we commit. My past projects with other platforms taught me that the devil is often in the details like egress fees or incremental indexing costs 😅
Thanks in advance for any real-world insights!
- Kev
Good luck getting a straight answer. Their pricing page is a masterpiece of obfuscation.
Your real cost won't be the 500k documents. It'll be the query volume and the inevitable storage for embeddings. They charge for both, of course. If your docs are large PDFs, prepare for a multiplier on those "records" they love to mention.
Performance is fine until you hit a concurrency limit, then you're negotiating a custom plan. That's where the real bill starts.
Your stack is too complicated.
Yeah, that's the exact wall I hit. Their trial's great, but the pricing gets fuzzy.
We're on the "Pro" plan, about 300k docs. Our bill is almost double what we guessed. The 500k "records" isn't just documents. If your PDFs get chunked for search, each chunk is a record. So you might have 500k *documents* but 2 million *records*.
Biggest cost for us was queries. The included amount on the plan was way too low for actual user traffic.
Are your PDFs mostly text, or do they have a lot of images? That embedding storage cost is real.
Still learning
You're right about chunking. That's the hidden multiplier. I've seen teams index large legal PDFs where each page is a chunk, blowing their record count by 10x.
Your query cost problem is common. Their plans include a paltry number of "query units". One complex filter with vector search can consume 5-10 units per request. You burn through the allowance in days.
For 500k source docs, you're looking at a custom Enterprise quote. The real metric to calculate is total chunk count + peak queries per second. Don't let them sell you on the source document number.
Trust, but verify