Amazon Comprehend
Amazon Comprehend scores 85/100 on xpay's Agent-Ready Index v2: partial pricing visibility; self-serve checkout is available; a public API exists with documentation; usage-based / per-token pricing is published; a free entry path exists.
Amazon Comprehend pricing
Sign up, get an API key, pay per call
Self-serve checkout with public pricing — an autonomous agent can complete the purchase without a sales call.
7 SKUs priced in mixed units (characters, MB, hour…)
| Model / SKU | Input | Output |
|---|---|---|
Standard NLP | $0.0001 / 100 characters | — |
Custom Classification (async) | $0.0005 / 100 characters | — |
Topic Modeling (first 100 MB) Flat rate for up to 100 MB | $1 / 100 MB | — |
Topic Modeling (additional MB) | $0.004 / MB | — |
Model Training Billed by the second | $3 / hour | — |
Model Management | $0.5 / month | — |
Inference Unit (IU) Per IU provisioned | $0.0005 / second | — |
How Amazon Comprehend scores on the agent-ready dimensions
Public pricing visibility
15 / 15
Self-serve checkout
15 / 15
Public API + docs
16 / 20
Usage-based / metered pricing
20 / 20
Free entry path
7 / 7
No sales call required
7 / 7
Card / wallet / x402 payment
5 / 6
Machine-readable bonus (MCP / OpenAPI)
0 / 10
Total
85 / 100
What it actually costs at typical scale
10,000 requests of 550 characters each (second year, no free tier)
$6/mo
Custom classification async + model training for one month
$25.1/mo
Can an autonomous agent transact here?
Can create account
Can enter payment method
Can upgrade plan
Can buy more usage
Can cancel online
Payment methods
Human approval recommended
Pricing transparency
full
Purchase motion
self_serve
Demo required
Quote required
Annual commit required
Auto-renewal mentioned
Cancellation policy visible
Refund policy visible
Invoice payment available
Purchase orders supported
MSA required
DPA available
Security review required
Trial
Available, no card
API & developer surface
API-first product
Public API docs
Sandbox present
Webhooks
MCP mentioned
Positioning & ecosystem
Buyer persona
Six-step check: can an agent actually buy from Amazon Comprehend?
Discover price
Select a plan
Pay per task
Avoid a sales call
API docs without auth
Estimate cost upfront
Pros and cons for AI agents
Observational summary written by xpay from the signals captured on 2026-05-06. Not a review of the product — only of its current pricing posture for agent buyers.- Pricing is publicly visible on an indexable page — agents can read tiers without scraping past auth.
- Every advertised tier is self-serve; no tier requires scheduling a call.
- A free tier exists, lowering the bar for an agent builder to prototype before committing budget.
- No per-unit price was advertised, so an agent has no way to estimate the cost of a single task.
- API documentation is gated or absent; an agent cannot inspect the integration surface without authentication.
- PPU requires account creation and an issued API key; a fully unauthenticated agent purchase is not yet supported.
- No /.well-known/ai-pricing.json or equivalent machine-readable pricing manifest is published — agents must rely on HTML scraping.
How Amazon Comprehend could lift its score
Add a per-unit price (e.g. $X / 1K calls) to the pricing page so an agent can compute its own cost before committing.
| pricing_visible | true |
| headline_phrasing | Amazon Comprehend pricing — Amazon Comprehend provides natural language processing, Personal Identifiable Information (PII) detection and redaction, Custom Classification and Entity detection, and topic modeling, enabling a broad range of applications that can analyze raw text, and with some APIs, document formats like PDF and Word. |
| tier_count | 3 |
| lowest_paid_entry_usd | 0.5 |
| free_tier | true |
| free_tier_terms | null |
| per_unit_price | $0.0005 per second |
| annual_required | false |
| self_serve_paid_tiers | 2 |
| sales_only_tiers | 0 |
| public_api_docs_url | https://builder.aws.com/ |
| api_docs_auth_walled | false |
| ai_pricing_json_present | false |
| agents_txt_present | false |
| anonymous_purchase_path | false |
| per_unit_classification | {"unit_class":"compute_time","confidence":"medium","agent_task_price_forecastable":"partial","rejected_reason":null,"evidence_text":"$0.0005 per second"} |
| usage_headline_present | false |
| custom_tier_present | true |
| agent_friendly | {"ai_pricing_json":false,"agents_txt":false,"llms_txt":false,"sitemap_xml":false,"mcp_server_card":false,"agent_skills_index":false,"x402_supported":false} |
View raw extracted page text →
Claim this scorecard & lift your score.
Get the breakdown of every signal we measured, the one-line fix, and how your peers in Natural Language Processing stack up.
Compare the agent-ready picks in Natural Language Processing.
If you're building or running an AI agent that needs to buy from this category, see who's scoring highest right now.
See agent-ready picksAmazon Comprehend
85
/ 100 (rubric v2.0)Natural Language Processing
Not advertised
3
Yes
$0.5/mo
$0.0005 per second
2 / 0
Public
No
Not published
2026-05-09
Pay-as-you-go API
Amazon Comprehend pricing starts at $0.0001 per 100 characters unit; a free tier of 50,000 units (5 M characters) is available; paid usage up to $0.004 per MB for topic modeling. Billed per usage.
Free tier (Free)
Discovery files and protocols
Side-channel signals — informational, not part of the score. Each protocol is independent; adoption signals the publisher is thinking about agent buyers.ai-pricing.json
agents.txt
llms.txt
sitemap.xml
MCP Server Card
Agent Skills
x402 / MPP
Compare Amazon Comprehend with peers
Highest-scoring companies in the same category that an AI agent might evaluate as an alternative.MCP-Builder.ai
83
Forefront
71
NLP-API
66
Ana By TextQL
61
Five
55
Powerdrill AI
49
Spark NLP
39
QnA Maker
26
Explore the Agent-Ready SaaS Index
25,480+ SaaS scored on agent-buyability — browse by readiness band, category, or jump straight to a pay-per-run tool on xpay.tools.