Recognize the question
Terms such as Data Residency, Human Approval, Audit Logs, RBAC or MCP are mapped to normalized fields.
Ask AgentenCode
Ask AgentenCode — by text or voice.
Ask a natural-language question about AI agents, capabilities, privacy, governance, protocols or changes. AgentenCode answers from the published evidence base and shows why a result matches — or why it does not.
130 profiles802 Trust signals1.235 Evidence claims79 Relations250 History events
Direct answer
An evidence search engine instead of a free-form hallucinating chatbot.
Ask AgentenCode recognizes agents, trust controls, protocols and time-based questions. Full matches must satisfy the recognized criteria. Near matches expose evidence gaps or documented conflicting values. Every answer remains traceable to scope, review date and primary source.
You can start immediately.
Examples: “Which agents have Data Residency and SAML SSO?”, “Does OpenAI Codex support MCP?” or “What changed for Copilot Studio since 2026-10-01?”
How it works
Natural-language question, structured evidence query.
The system does not simulate open-ended chat. It maps verifiable questions to the structured dataset.
Connect evidence
Agent, governance, evidence and relations data are evaluated together.
Separate matches
Full matches satisfy all recognized criteria; evidence gaps remain visible.
Show evidence
Scope, review date and primary source make the result traceable.
Example questions
Questions that map well to structured evidence.
The more specific the verifiable criteria, the more robust the result.
„Which agents have Data Residency, Human Approval and Audit Logs?“
AND query across multiple trust controls„Which agents support MCP?“
Relations and evidence query„What is documented for OpenAI Codex on RBAC and Human Approval?“
Profile plus two governance fields„Which agents document that customer data is not used for training?“
documented_false can be a positive selection criterion hereLimits
What Ask AgentenCode does not do.
The engine does not invent answers outside the published data. It does not make legal or compliance judgments, generate hidden product scores, or interpret Unknown as No.
The current version intentionally covers a limited set of normalized terms and relations. It is less flexible than a generative chat system, but its answers are reproducible and traceable to concrete evidence.
AgentenCode tools
Ask, discover, check and compare — without mixing tool purposes.
Ask AgentenCode answers evidence-backed questions. Specialized tools handle discovery, requirement checks, fit and direct comparisons.
01
Agent Finder
Shortlist AI agents by task, Audience and provider.
Find agents →02Check
Check concrete requirements against evidence, trust and open risks.
Check requirements →03Fit
Assess whether an agent, copilot, automation or hybrid approach fits the process.
Check fit →04Compare
Compare multiple agents side by side at field level.
Open comparison →FAQ
Frequently asked questions
Does Ask AgentenCode use an external language model?
No. The current query is deterministic and works only on the published structured dataset.
Can I use Ask AgentenCode by voice?
Yes, if your browser supports web speech recognition. AgentenCode itself does not store audio recordings. Depending on the browser or operating system, recognition may use a service from the relevant platform provider.
Why does a question sometimes return few matches?
Multiple criteria are interpreted as joint requirements. A full match needs suitable evidence for every recognized criterion.
What happens when evidence is missing?
Missing evidence remains Unknown. It is not interpreted as No.
Which sources are shown?
Where available, the answer shows the underlying primary source together with scope and review date.