Head-to-head
Akamai vs HUMAN Security
In short
Akamai is listed by OpenAI as recognising ChatGPT's Cloud browser under the bot name "ChatGPT Agent"; HUMAN lists the same agent in Known Bots & Crawlers and AgenticTrust permissions. On per-purpose policy the rubric grades Akamai partial and HUMAN Security yes, and on commerce agents the rubric grades Akamai yes and HUMAN Security not published.
The sourced difference
Akamai is listed by OpenAI as recognising ChatGPT's Cloud browser under the bot name "ChatGPT Agent"; HUMAN lists the same agent in Known Bots & Crawlers and AgenticTrust permissions. Source, read 2026-11-04.
Side by side
| Question | Akamai | HUMAN Security |
|---|---|---|
| Tells a verified agent from a spoof | Yes Web Bot Auth (RFC 9421) at the edge, verified bot list incl. AI bots; Visa TAP and Skyfire/Experian KYA via the agentic framework |
Yes per-request cryptographic keys (e.g. ChatGPT's published key), plus behavior and private signals; Web Bot Auth not named on pages read |
| Per-purpose policy (train, search, answer, act) | Partial allow, challenge, delay or block per verified bot or category; no published search versus training switch (checked 2026-11-04) |
Yes per-action categories (content, engage, account creation, checkout, rate limit) by trust level |
| Charging for access | Partial through TollBit and Skyfire partners, no native billing |
Partial through scraping monetization partners (TollBit logo on product page); no native charging found |
| Lets signed commerce agents through | Yes Visa Trusted Agent Protocol integration, Skyfire and Experian KYA |
Not published Visa TAP or Mastercard Agent Pay support (checked 2026-11-04) |
| Cost | no public price; sales-led | no public price; sales-led |
| Reporting | Yes TrafficPeak and web security analytics |
Yes user and session insights on agent actions, monthly State of Agentic Traffic benchmark |
| Where it runs | edge: verification and enforcement on Akamai's edge network | platform: Sightline integrates with the customer's stack; exact deployment points not confirmed on pages read |
| Points, rank | 8 of 10, rank 1 (tied) | 7 of 10, rank 5 |
| Price read 2026-11-04 | no public price; sales-led (checked 2026-11-04) | no public price; sales-led (checked 2026-11-04) |
Verdicts by reader
- Platform engineer
- Both verify the signature automatically, per OpenAI.
- Security engineer
- HUMAN adds per-action permissions; Akamai adds allow, challenge, delay or block actions.
- Publisher operations
- Akamai can route bots to TollBit and Skyfire; HUMAN shows TollBit as a partner.
What each does not do
Akamai
- No native per-crawl billing; monetization runs through TollBit or Skyfire (Akamai blog and partnership release).
- The June 2026 framework release gives no availability dates or prices (press release read 2026-11-04).
- No published control that separates search crawling from model training as purposes (pages read 2026-11-04).
HUMAN Security
- Medium-trust agents lack cryptographic verification, so HUMAN itself warns of false negatives (Trust Levels blog, 2026-09-02).
- No native pay-per-crawl; monetization is through partner tools (AgenticTrust page).
- No published price list (package plans page).
More: alternatives to Akamai, alternatives to HUMAN Security.
Sources
- ChatGPT Work's Cloud browser allowlisting
- akamai.com: akamai unveils agentic security framework to power trusted ai driven interactions and commerce
- humansecurity.com: package plans
- akamai.com: redefine trust web bot authentication
- humansecurity.com: evaluating ai agent trust humans agent trust levels
- akamai.com: no free crawls akamai tollbit and skyfire turn traffic into revenue
- humansecurity.com: agentic ai
- akamai.com: akamai and visa join forces to secure the next era of agentic commerce
Last reviewed 2026-11-04