Pages
Conversations
Make requests to our ontology models and generate graphs from text
| TOKENS | PRICE PER 1,000 |
|---|---|
| 0-100k | FREE |
| 100,001-1,000,000 | $0.0300 |
| 1,000,001-10,000,000 | $0.0150 |
| 10,000,001-100,000,000 | $0.0125 |
| 100,000,001-1,000,000,000 | $0.0100 |
| 1,000,000,001+ | $0.0050 |
Tokens are common sequences of characters found in text. The below text is tokenized. Our system uses the tiktoken library.
13
Store vector and graph data in at least three regions worldwide
| STORAGE SIZE | PRICE PER GB / DAY |
|---|---|
| 0 - 50 MB | FREE |
| 50 MB - 1 GB | $1.3142 |
| 1 GB - 5 GB | $1.1499 |
| 5 GB - 10 GB | $0.9856 |
| 10 GB - 25 GB | $0.8214 |
| 25 GB - 50 GB | $0.5750 |
| 50 GB + | $0.3285 |
Make requests to the retrieval API. This does not include the ingestion API.
| REQUESTS | PRICE PER REQUEST |
|---|---|
| 0 - 100 | FREE |
| 101 - 1,000 | $0.0300 |
| 1,001 - 10,000 | $0.0150 |
| 10,001 - 100,000 | $0.0125 |
| 100,001 - 1,000,000 | $0.0100 |
| 1,000,001+ | $0.0050 |
Below are some examples of how much you might expect to pay for different use cases.
Julien is creating a voice agent for his company. He has a website containing all of the company's products, totalling 1,000 pages and their descriptions, which are about 4 paragraphs in length. He wants the voice agent to be able to answer the questions that customers often use the 1-800 number to ask.
The below workflow is most suitable for Julien, and it is priced accordingly.
FREE
1,000 pages * 20 facts per page = 20,000 facts
15 tokens per fact = 300,000 tokens
300,000 tokens - 100,000 free = 200,000 @ $0.03 / 1,000 = $9.00 USD
200,000 facts * 100 bytes per fact = 20,000,000 bytes = 20 MB
Storage - 100 MB free = FREE
About 20 people call Julien's 1-800 number every day, so about 600 calls per month. (well done Julien!) On average, each call lasts 7 minutes and throughout that, there are 5 API calls per minute. Let's estimate 35 API calls per call.
600 calls * 35 API calls per call = 21,000 API calls
21,000 API calls - 100 free = 20,900 @ $0.0125 = $261.25 USD
Will Smith owns a voice AI company with 100 customers. Each customer has a specific use case, for example, they work in the medical products industry and want their customers to be able to call up, enquire about products, and make appointments.
Each customer's information spans approximately 50 pages, with 500 words per page. Will processes all customer data at once and then manages ongoing monthly costs.
100 customers * 50 pages * 500 words = 2,500,000 words
Average 1.5 tokens per word = 3,750,000 tokens
Coreference resolution is roughly equal to initial content tokens = 3,750,000 tokens
3,750,000 tokens - 100,000 free = 3,650,000 @ $0.0150 / 1,000 = $54.75 USD
3,750,000 tokens x ~ 20 (since we only count output tokens) = 75,000,000 tokens
75,000,000 tokens @ $0.0150 / 1,000 = $937.50. USD
2.5M words * 4 bytes per token = = ~ 600 MB
600 MB - 100 MB free = 600 MB = 0.5 GB @ $0.50 per GB = $19.71 USD per month
100 customers * 5 calls per day * 30 days = 15,000 calls per month
Average 5 minutes per call * 5 API calls per minute = 25 API calls per call
15,000 calls * 25 API calls = 375,000 API calls per month
375,000 API calls - 100 free = 374,900 @ $0.0125 = $4,686.25 USD per month
