OBSERVED BEHAVIOR INDEX
Descriptive activity only. Creditworthiness and financial risk are not assessed.
Invited users can sign in to the app for live checks. Explore the illustrative sample here, or open the private app.
Open private appONCHAIN OBSERVATIONS
Explore activity, execution and history—with the evidence in view.
For a protected service, enter your MyCryptoWalletScore access token. It stays in memory until you disconnect or close this page. Never enter a seed phrase, private key or chain-provider key.
Local API access is used by default.
YOUR NEXT LOOKUP
Choose a network to see what we can observe.
Reports keep missing evidence visible.
OBSERVED BEHAVIOR INDEX
Descriptive activity only. Creditworthiness and financial risk are not assessed.
One address, one network. Matching addresses across chains do not establish shared ownership.
KNOW THE BOUNDARIES
Implemented adapters have different scopes. Live availability depends on the provider.
| Network | Provider access | What is observed |
|---|
ONE EVIDENCE SERVICE
The same observations for your AI tools and your applications.
Remote MCP tools are prepared locally. Public connection needs hosting, configured OAuth and platform verification.
OpenAI integration guideLocal stdio MCP is tested. Cloud connectors need a reachable HTTPS endpoint and a verified sign-in flow.
Claude connector guideA consuming application can configure remote MCP and securely supply its MyCryptoWalletScore access token.
Groq integration guideWallet assessment · Side-by-side comparison · Supplied snapshot replay · Methodology · Chain catalog · Current staking · Bounded staking history · Participation replay
The local website uses the REST API. AI clients use MCP. No public installation or publication has occurred.
TRANSPARENT BY DESIGN
An experimental description of observed activity, with explicit evidence requirements.
Uses the exact first observation and a logarithmic age scale, capped at 730 days.
The share of fully covered calendar months with successful activity. Unknown outcomes withhold this category; sampled fragments do not count as complete months.
Successful observed executions divided by all observed executions. Every execution must have a known status.
An overall index requires at least 10 transactions, 30 days of observed history, complete supported history and all three categories. Otherwise the index is unavailable.
Higher values describe older history, more regular activity and more completed executions. They do not establish better economic behavior.
No credit rating, financial risk, profits, sanctions screening, wallet ownership or unique-person assessment.
Model: observed-behavior-0.2.0
TIME, WITH EVIDENCE
Current positions and duration supported by available records. Recurring deposits alone do not prove staking.
Ethereum: direct stETH / wstETH exposure, which is transferable and unlocked. Solana: current stake accounts for the withdrawal authority; historical duration is pending. Other staking protocols are not covered yet.
Protected service? Set your API access token in Wallet explorer first.
Time is counted once across concurrent positions. Milestones do not change the behavior index, prove loyalty or grant rewards.
THE PATH TO A PARTNER PILOT
A working local evidence service, with clear gates before public use.
| Section | Status | Next evidence needed |
|---|---|---|
| Evidence API + AI tools | Built locally | Eight read-only tools; shared service, access controls and bounded requests. |
| Wallet behavior + chains | Implemented with limits | 47 address routes; live access and coverage vary. The 0–100 index describes activity. |
| Current staking | Built locally | Native Solana and direct Ethereum Lido holdings. Broader protocols need their own adapters. |
| Ethereum participation history | Bounded prototype | Archive access, live independent comparison and wider protocol event validation. |
| Solana participation history | Active-epoch audit built; validation pending | Closed-epoch rewards and stable authority tails support bounded active-day lower bounds. Archive access, active balances and independent certification remain pending. |
| Milestone reports + website | Built locally | Unknown evidence stays visible; synthetic candidates are labeled. |
| Partner accuracy pilot | Pending partner data | Reviewed examples, known histories and independently measured error rates. |
| Public AI connectors + scale | Deployment package prepared | Hosting, domain, identity provider, platform connection tests and load tests with live providers. |
| Partner benefits | Pending agreements | Partners define opt-in benefits and eligibility. No rewards are promised. |
A SMALL EVIDENCE BETA
Explore observed wallet activity and supported staking evidence. Every report keeps its limits visible.
Mainnet wallet observations, current Solana staking accounts, Ethereum Lido holdings and bounded history when records are available. The experimental behavior index runs from 0 to 100.
47 address routes have different provider requirements. A supported route does not mean every token, staking protocol or historical event is covered.
Unknown means the available records do not establish the answer. It does not mean a bad wallet. Liquid staking tokens are transferable exposure, not proof that money is locked.
Reports do not establish financial credit, repayment, personal identity, customer loyalty or reward eligibility. Verified staking milestones remain disabled.
This site introduces the private trial. Live checks run in the separate signed-in app; public launch remains pending.
KNOW WHERE YOUR DATA GOES
Private trial notice. The public website is hosted on Vercel. The app, authentication and shared request quotas are hosted on Zoho Catalyst. Public support and retention details are still being finalised.
We receive the network and public address you enter, and any snapshot or replay inputs you supply. Public addresses can identify people. We send requested addresses/account identifiers to configured blockchain providers to fetch evidence.
AI clients receive reports when you use their connected tools. Their own policies apply. MyCryptoWalletScore does not call an AI model to calculate your report.
Reports may be cached in memory or the configured database. Default cache validity is five minutes, with bounded storage. Expiry is not a guarantee of immediate erasure from backups or filesystem history. Request quotas use hashed identifiers.
Application logs contain request ID, method, route, status and elapsed time; they omit query strings, bodies and authorization headers. Hosting, identity and chain providers may keep their own logs.
This website does not collect login passwords or access tokens. The separate app uses a Zoho-managed browser login session. No third-party analytics script is included.
The operator must supply a public contact, hosting/provider information, retention settings and support/deletion process. These details are pending; no public privacy commitment or deletion service is being represented as complete.
Never enter a seed phrase, private key, exchange password or chain-provider secret.
START WITH EVIDENCE
A provider may be unavailable, the requested history may be incomplete, or the evidence may fail validation. The service withholds conclusions in those cases. Repeated queries cannot create missing evidence.
The sample uses synthetic supplied intervals. Its milestones are unverified candidates, not real wallet badges or rewards.
For a protected deployment, use a MyCryptoWalletScore access token from the operator in Wallet explorer. It stays in page memory. Never enter a blockchain provider key.
Public support contact is awaiting operator configuration. When reporting an issue, include the support reference shown with the error, network and observation time; omit all secrets. Long lookups can be cancelled with the Cancel lookup button. Try a sample to explore without a live provider request.
TRY THE FULL EXPERIENCE
Try these six tasks on this computer. This page does not submit feedback or record participants.
Is this a financial credit rating? Does unknown mean a bad wallet? Does liquid token exposure prove locked funds? Is the synthetic sample independently verified?
The intended answer to each is no. Explain your reasoning before checking with the reviewer.
Tell the reviewer which task was confusing, what you expected and what happened. Include an error's support reference if relevant. Do not include credentials, personal details or private wallet information.
The reviewer records task outcomes locally using the trial kit. Automated checks and agent walkthroughs are kept separate from human feedback.
The private app is hosted on Zoho Catalyst and is available to invited users. Completing sign-in alone does not complete the user trial; task outcomes and feedback are still being reviewed.