GAIP Agent Observatory
State of public AI agents, October 2026
What GAIP observed across public MCP servers and A2A agents: how many answered, which protocol versions they report, how many ask for credentials and which discovery standards hosts publish. Aggregate facts only; no party is scored or ranked.
GAIP Agent Web Index — October 2026 (provisional: the month has not ended)
Declared versus observed: what public AI agents declared and what changed this month. Headline figures count every public agent or host GAIP read this month (22735 agents watched). Beside each: the frozen panel AGENTS-COHORT-2026-10, 4157 public MCP servers and A2A agents on 3676 hosts. Registry size: the Official MCP Registry listed 44477 server entries in GAIP's last complete walk this month (completed 2026-10-17), 27634 of them with a remote endpoint GAIP can read; GAIP watched 21873 of its servers and read 21830 of them in the last 7 days. Aggregates only: a count is shown only when it and its total are at least 5; no host, agent or caller is named; facts are as observed by GAIP at the stated time and assess no one. A month is final only after it ends; until then it is provisional.
| Figure | All watched agents | Out of | Panel AGENTS-COHORT-2026-10 | Out of |
|---|---|---|---|---|
| Agents whose declared tools changed | 2322 of 22735 (10%) | every public MCP server and A2A agent GAIP watched when this copy was built (all sources; hosts an outside caller asked about and GAIP's own agents left out); the agents counted are among them | 421 of 4157 (10%) | every member of the frozen panel AGENTS-COHORT-2026-10 |
| Hosts whose declared terms, limits or prices changed | 708 of 9812 (7%) | every public host whose declared terms GAIP read this month (hosts an outside caller asked about and rows under a method version with a known defect left out) | 322 of 1786 (18%) | panel hosts whose declared terms GAIP read this month (agent_terms rows) |
| Agent cards carrying a signature | 52 of 713 (7%) | every public host whose A2A agent card GAIP read this month (latest read; hosts an outside caller asked about and rows under a method version with a known defect left out) | 36 of 482 (7%) | panel hosts whose A2A agent card GAIP read this month (latest read) |
| Most frequent failure types diagnosed for outside callers | too few diagnoses to show (under 5 per type) | front-door gaip_diagnose answers to outside callers this month | ||
| Hosts carrying any AI-preference signal | 1893 of 20020 (9%) | every agent host whose robots.txt or responses GAIP read this month (latest read; hosts an outside caller asked about and rows under a method version with a known defect left out) | 339 of 3486 (10%) | panel hosts whose robots.txt or responses GAIP read this month (latest read) |
| Agent prices (x402 Bazaar listings) | not collected yet (switched off until the founder confirms the source terms) | |||
| ERC-8004 agent registration files | not collected yet (switched off until the founder confirms the source terms) | |||
| Reachability and response times (the 30 days to 17 October 2026) | 97.6% of 86350 visits reachable; 94.0% of 18777 agents reachable on at least 95% of visits; response time p50 229 ms, p95 1129 ms | |||
| Reachability and response times, panel AGENTS-COHORT-2026-10 | 97.8% of 52905 visits reachable; 92.2% of 3785 agents reachable on at least 95% of visits; response time p50 226 ms, p95 1119 ms | |||
What agents ask for, depend on and who runs them (latest frozen week, all watched agents): Agents whose tools ask for personal, payment or secret data (by field category): 2068 of 9041 (23%); Agents asking for payment card fields: 6 of 9041 (0%); Agents asking for email fields: 1172 of 9041 (13%); Agents asking for phone fields: 469 of 9041 (5%); Agents asking for credentials fields: 585 of 9041 (6%); Agents asking for precise location fields: 376 of 9041 (4%); Agents whose declarations point to an outside domain: 9765 of 20540 (48%); Declared operators: 15755; agents run by the 10 largest: 1138 of 20450 (6%).
Cite as: GAIP Agent Web Index, October 2026, gaipagents.com/state-of-agents?month=2026-10, CC BY 4.0. Machine-readable: /v1/free/agent-web-index. How to reproduce each figure: /v1/free/agent-web-index/queries.
Reachability and response times: the 30 days to 17 October 2026 (provisional)
Aggregate facts observed by GAIP at the stated times: how often public agents answered and how fast. Not an assessment of any party; no agent or host is named. Cells under 5 agents are withheld.
| Cohort | Protocol | Figures |
|---|---|---|
| All public agents | MCP and A2A | 97.6% of 86350 visits reachable; 94.0% of 18777 agents reachable on at least 95% of visits; response time p50 229 ms, p95 1129 ms |
| All public agents | MCP | 97.5% of 78338 visits reachable; 94.0% of 17944 agents reachable on at least 95% of visits; response time p50 240 ms, p95 1159 ms |
| All public agents | A2A | 98.6% of 8012 visits reachable; 93.9% of 833 agents reachable on at least 95% of visits; response time p50 175 ms, p95 788 ms |
| Panel AGENTS-COHORT-2026-10 | MCP and A2A | 97.8% of 52905 visits reachable; 92.2% of 3785 agents reachable on at least 95% of visits; response time p50 226 ms, p95 1119 ms |
| Panel AGENTS-COHORT-2026-10 | MCP | 97.6% of 45563 visits reachable; 91.6% of 3324 agents reachable on at least 95% of visits; response time p50 243 ms, p95 1162 ms |
| Panel AGENTS-COHORT-2026-10 | A2A | 99.0% of 7342 visits reachable; 96.8% of 461 agents reachable on at least 95% of visits; response time p50 173 ms, p95 794 ms |
Cite as: GAIP Agent Reliability Index, the 30 days to 17 October 2026, gaipagents.com/state-of-agents, CC BY 4.0. Downloads: CSV · JSON.
Collection record
When GAIP's Observatory was running, from its own daily ledger (since 2026-10-04; 13 days frozen and hash-chained, head 4dbbcccd48a13636b69fc3490e2535def01d9e20a443aa17b761d6701df3c586).
Last 30 days: 13 days, 1336 ticks, 11 days fully covered; a tick started in 98.6% of 15-minute windows. Gaps (no tick start for more than 30 minutes): 1.
- 2026-10-09T16:34:12Z to 2026-10-09T17:32:32Z (58 min)
Known gaps (register, append-only):
- 2026-10-02T23:16:00Z to 2026-10-03T07:52:15Z (times partly estimated): Observatory observations: every tick claimed and observed 0 targets (tick lines showed budget_exhausted or TimeoutError), so no status observations, manifest changes or capability records were written. Cause: Registry walks read up to 20 pages from one host while the outbound gate allows 10 requests a minute per host; the 11th page raised _BudgetExhausted out of the walk, the cursor was never saved and the cycle returned before observing, on every tick (the refresh livelock). Fixed by PR #824.
- 2026-10-01T21:09:04Z to 2026-10-03T10:02:34Z (times partly estimated): About 75% (estimate) of shop-host AI-permission facts (ai_permissions dataset, shop hosts) were evicted from an in-memory queue before they were written; the rest were written. Cause: The in-memory buffer evicted entries waiting for an extra read before they were written; the shop panel did not read tdmrep.json itself. Fixed by PR #826.
- 2026-10-03T16:55:00Z to 2026-10-03T17:14:00Z (times partly estimated): One Observatory tick observed nothing: the tick right after the data5 deploy ended with DatabaseError. Cause: Deploy overlap during the PR #843 deploy (as recorded in the handover); no code fix was needed, the next tick was clean.
- 2026-09-28T00:00:00Z to 2026-10-01T19:30:20Z (times partly estimated): No independent daily timestamp (Sigstore/Rekor anchor) of the history log's chain head was kept for 28, 29 and 30 Sep 2026 or the morning of 1 Oct. The history log itself kept every entry; the 1 Oct anchor's chain head covers every entry before it, so those entries are dated no later than 1 Oct by an outside witness, not day by day. Cause: The anchor workflow's runs on 28 Sep failed twice and were then cancelled; it ran on GitHub Actions only on a push or by hand until the COO schedule dispatcher began running it daily on 1 Oct 2026 (moat moves), and dated anchor files began the same day (record durability).
Method: collection ledger and known gaps. Downloads: collection-ledger.csv · collection-ledger.json.
Week 2026-W41 (frozen)
| Public AI agents and MCP servers in the Observatory | 20673 |
|---|---|
| Observed this week | 20603 |
| Answered at least once | 15345 of 20603 (74%) |
| Median response time (ms) | 231 |
| MCP servers that asked for credentials | 4322 of 19755 (22%) |
| MCP servers listing tools without credentials | 10254 of 19755 (52%) |
| New this week | 2414 |
| Published a new manifest version | 5530 |
| Hosts probed for discovery standards | 0 |
| Protocol versions reported | 0.2.5: 5, 0.3: 22, 0.3.0: 344, 1.0: 90, 1.0.0: 12, 2024-11-05: 507, 2025-03-26: 474, 2025-06-18: 1533, 2025-11-25: 6587, 2026-07-28: 123, other_categories_under_5_each: 14 |
This week so far (2026-W42, provisional)
| Public AI agents and MCP servers in the Observatory | 22982 |
|---|---|
| Observed this week | 17829 |
| Answered at least once | 13367 of 17829 (75%) |
| Median response time (ms) | 227 |
| MCP servers that asked for credentials | 3863 of 16959 (23%) |
| MCP servers listing tools without credentials | 8714 of 16959 (51%) |
| New this week | 2311 |
| Published a new manifest version | 3786 |
| Hosts probed for discovery standards | 0 |
| Protocol versions reported | 0.2.5: 5, 0.3: 22, 0.3.0: 337, 1.0: 101, 1.0.0: 12, 2024-11-05: 385, 2025-03-26: 404, 2025-06-18: 1315, 2025-11-25: 5536, 2026-07-28: 116, other_categories_under_5_each: 13 |
Delivery outcomes, 2026-10
What happened to calls outside callers made to GAIP's own public agents this month, and the verdicts of the Witness delivery checks they asked for. Counts about GAIP's own service only; no caller or third party is named. A share is shown only when both numbers are at least 5.
Outside calls: 1849
| Error | 75 of 1849 (4%) |
|---|---|
| Blocked | 599 of 1849 (32%) |
| Completed | 909 of 1849 (49%) |
| Needed more input | 266 of 1849 (14%) |
| GAIP agent | Outside calls | Completed | Needed more input | Blocked | Error |
|---|---|---|---|---|---|
| witness | 88 | 30 | 0 | 54 | 0 |
| front-door | 3 | 0 | 0 | 0 | 0 |
| unattributed | 1305 | 786 | 266 | 182 | 71 |
| trust-assurance | 36 | 26 | 0 | 10 | 0 |
| listing-observer | 3 | 0 | 0 | 0 | 0 |
| integration-repair | 6 | 0 | 0 | 5 | 0 |
| procurement-verify | 285 | 0 | 0 | 284 | 0 |
| integration-protocol | 123 | 64 | 0 | 59 | 0 |
Witness delivery checks: 0
| Met the registered terms | 0 of None |
|---|---|
| Did not meet the registered terms | 0 of None |
| Could not be decided | 0 of None |
Outcome ledger, 2026-10
GAIP's own outcome ledger: one row per month of what happened in agent work GAIP saw, for outside callers only, frozen after the month ends and hash-chained to the month before (never changed or deleted; the chain head is signed in the daily Sigstore anchor). Counts only; categories under 5 are folded together; no caller or third party is named.
This month's row is frozen once the month has ended (from 1 November 2026).
Agentic-commerce shop panel, week 2026-W41
A fixed panel of 392 online shops (cohort shops-2026-10), read once a week: which agent-commerce discovery documents they publish and whether product prices and stock agree across a page, its structured data and the shop's UCP catalog. No shop is named; shares are shown only when both numbers are at least 5.
Shops observed: 360
| UCP profile (/.well-known/ucp) | 247 of 360 (69%) |
|---|---|
| /.well-known/x402 | 0 of 360 |
| llms.txt | 235 of 360 (65%) |
| A2A agent card | 0 of 360 |
| x402 payment request on the declared UCP endpoint | 0 of None |
| Products read | 269 |
| Price agreement across sources | SAME: 187, NOT OBSERVED: 78, other categories under 5 each: 4 |
| Stock agreement across sources | SAME: 111, NOT OBSERVED: 156, other categories under 5 each: 2 |
What agents changed this week (2026-W41, frozen)
Changes to what public MCP servers and A2A agents declare, by kind. A change of the protocol version an agent answered with is not counted as the agent changing what it offers; it is shown on its own line. A substantive change is silent when no declared version (registry or card version, server version) moved that week. Observed at the stated time; no agent is named. Kinds with fewer than 5 agents show as under 5.
| Agents | All watched | Panel AGENTS-COHORT-2026-10 |
|---|---|---|
| Agents watched | 20673 | 4145 |
| Agents that changed what they declare (protocol-version-only changes left out) | 2156 of 20673 (10%) | 520 of 4145 (13%) |
| Agents whose only change was the protocol version they answered with | 0 of 20673 | 0 of 4145 |
| Agents with a substantive change (tools, descriptions, formats, labels, sign-in, terms, endpoint) | 1510 of 20673 (7%) | 328 of 4145 (8%) |
| Of those, changed without any declared version change (silent) | 994 of 1510 (66%) | 189 of 328 (58%) |
| Tools added | 519 | 119 |
| Tools removed | 124 | 31 |
| Tool descriptions re-worded | 1166 | 256 |
| Input or output format changed | 973 | 182 |
| Behaviour labels changed (read-only, destructive and similar hints) | 166 | 44 |
| Sign-in requirements changed | 19 | 9 |
| Terms or prices changed | under 5 | under 5 |
| Endpoint moved to another host | under 5 | under 5 |
| Declared version changed | 1141 | 320 |
| Protocol version changed | 62 | 6 |
| Tools per MCP server | 0: 5, 1: 552, 2-5: 3624, 6-10: 2641, 11-20: 1438, 21-50: 905, 51-100: 183, 101 or more: 33 |
|---|---|
| Tools marked read-only (readOnlyHint) | 57684 of 95314 (61%) |
| Tools marked destructive (destructiveHint) | 47377 of 95314 (50%) |
| MCP servers declaring an output format on a tool | 2803 of 9376 (30%) |
| MCP sign-in asked for in the registry entry | other header: 231, none declared: 12660, authorization header: 1970, api key or token header: 560 |
| A2A sign-in scheme types | http: 59, x402: 19, apiKey: 89, oauth2: 12, httpAuth: 30, none declared: 49, other categories under 5 each: 7 |
| A2A cards mentioning x402 | 245 of 753 (33%) |
| Registry entries this week (new, changed, removed) | 2414, 1804, 142 |
Long-run records, week 2026-W41
Four dated, hash-chained records GAIP keeps for the long run: the public keys agent hosts publish, what sites tell AI systems they may do with their content, how AI clients behave at GAIP (by client family only) and what agent services promise and charge; and the public npm and PyPI facts of MCP server packages with the security advisories OSV publishes for them (maintainers counted, never named). No host, package or client is named; shares are shown only when both numbers are at least 5.
| Agent hosts with a key history this week | 4152 |
|---|---|
| Agent cards carrying a signature | 47 of 642 (7%) |
| Agent hosts publishing /.well-known/jwks.json | 72 of 1506 (5%) |
| Panel shops blocking at least one AI crawler from the home page | 14 of 373 (4%) |
| Panel shops with a Content-Signal ai-train=no | 2 of 373 |
| Panel shops publishing a TDMRep reservation file | 0 of 332 |
| Agent hosts declaring terms, limits or prices | 8271 |
| Client families seen at GAIP (5+ requests) | 8 |
| MCP server packages (npm, PyPI) recorded this week | 4844 |
| Packages whose latest version declares install scripts | 179 of 4823 (4%) |
| Packages whose latest version has provenance | 1359 of 4818 (28%) |
| Packages with a published advisory affecting the latest version | 0 of 4844 |
AI-preference signals, 2026-10
How many hosts carried each machine-readable AI-preference signal when GAIP last read them this month: agent hosts (the Observatory's MCP servers and agents) and the fixed shop panel. Observed at the stated time; no host is named and nothing here assesses a site. A count is shown only when it and its total are at least 5. Parser: aipref-signals (MIT), following draft-ietf-aipref-vocab-08 (14 Sep 2026).
| Signal | Agent hosts | Panel shops |
|---|---|---|
| IETF AI preferences (Content-Usage) | 22 of 20019 (0%) | under 5 |
| IETF AI preferences: training refused | 8 of 20019 (0%) | under 5 |
| RSL licence link | 33 of 20019 (0%) | under 5 |
| W3C TDMRep file (tdmrep.json) | under 5 | under 5 |
| W3C TDMRep header | 29 of 20019 (0%) | 5 of 9106 (0%) |
| Content Signals | 1863 of 20019 (9%) | 88 of 9106 (1%) |
| Content Signals: ai-train=no | 541 of 20019 (3%) | 59 of 9106 (1%) |
| Any of these | 1893 of 20019 (9%) | 95 of 9106 (1%) |
What agents ask for, depend on and who runs them, week 2026-W41
Derived from what public MCP servers and A2A agents declared: the categories of data their tools' input fields ask for (counted by field category, never field names or text), the outside domains their declarations point to (well-known services named, others only counted) and how concentrated their declared operators are (no operator named). Not an assessment of any agent or operator.
- Agents whose tools ask for personal, payment or secret data (by field category): 2068 of 9041 (23%)
- Agents asking for payment card fields: 6 of 9041 (0%)
- Agents asking for email fields: 1172 of 9041 (13%)
- Agents asking for phone fields: 469 of 9041 (5%)
- Agents asking for credentials fields: 585 of 9041 (6%)
- Agents asking for precise location fields: 376 of 9041 (4%)
- Agents whose declarations point to an outside domain: 9765 of 20540 (48%)
- Most shared outside services: github.com (9502 of 20540 (46%)), gitlab.com (51 of 20540 (0%)), supabase.co (8 of 20540 (0%))
- Declared operators: 15755; agents run by the 10 largest: 1138 of 20450 (6%)
Cite this
GAIP Agent Observatory, State of public AI agents, October 2026. https://www.gaipagents.com/state-of-agents?month=2026-10 (weekly snapshots 2026-W41; latest SHA-256 0c9b250ad662e1d3e835ba4a5ac77f5d27aabdb0032fca10b2cc61bee4432781) Capability change-log head SHA-256 cccb299b02cb1f015b31c6792123e85a5ab6f978e9cd952a22ecbd3b8e83ab41 (34521 entries).
Machine-readable: /v1/free/state-of-agents. Weekly snapshots, hash-chained: /v1/free/ecosystem/snapshots. Method: /methodology. Aggregate figures are licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/): reuse them freely with credit to the GAIP Agent Observatory and the citation shown. Founder decision #13, 1 Oct 2026.