Check the source.
Read source references, observation times, available slot and finality information, and coverage notes alongside the result. Missing data is marked partial or unavailable.
“Show me the transactions behind this finding.”
Follow the money.
Keep the evidence.
Ask where a wallet’s funds came from, what changed around a token and which activity deserves a closer look. Give your agent the sources, the limits and the next question.
The foundation
Answers you can inspect.
These capabilities are implemented in the service. Access and available data depend on its configuration and your permissions. This page describes them; it does not run an investigation.
Trace the funding.
Start a funding investigation with a history window and bounded depth. Retrieve its result or cancel the job. Calls, observations and runtime have limits.
“Trace this wallet’s recent funding and show where the trail stops.”
Read wallet and token activity.
Inspect wallet activity, transaction facts, token supply, authorities and largest token accounts. Label and market enrichment depend on the configured providers and access tier.
“What can we establish about this token from the available records?”
Watch for changes.
Set event thresholds and cooldowns. With a configured event feed, alerts follow confirmed, finalized and retracted source events instead of treating every observation as permanent.
“Watch transfers above this amount. Keep retractions visible.”
Keep unknown cost basis visible.
The FIFO engine uses exact integer amounts and separates unknown lots. Available transaction classification and price coverage determine how much a wallet’s result can explain.
“Which parts of this wallet’s P&L can the records support?”
Give each agent the tools it needs.
Authenticated permissions control tool access. Configured holder tiers increase allowances; burn credits admit deeper jobs. The service does not sign or submit wallet transactions.
“Show my current tier, research limits and credit balance.”
Planned additions
Turn separate observations into an investigation.
The next step is to connect discovery, history and monitoring around a question you can keep following. The workflows below are planned, not available tools.
Find the buyers behind the pattern.
Discover wallets that repeatedly bought before a trader you follow. Compare their timing and funding paths, with possible relationships labelled separately from confirmed facts.
“Are these five early buyers independent, or is there evidence of shared funding?”
Follow a question between conversations.
Save a research objective and the wallets it concerns. Combine multiple events into a signal, then receive the evidence when the rule matches.
“Tell me when three wallets in this group buy the same new token.”
Open the case. See what changed.
Bring deployer history, funding routes, holder concentration and exits into one dossier. Compare saved investigations to see new evidence, revised conclusions and unresolved gaps.
“What changed in this investigation since yesterday?”
Follow the launch as it unfolds.
Put early buying, liquidity changes, distribution and deployer movements on one timeline. Return to the events behind a change instead of relying on a single risk score.
“Did ownership become more concentrated before the liquidity changed?”
Compare behaviour across tokens.
Review entries, exits and repeated associations across a wallet’s history. Show the observed sample and missing coverage beside each reputation label.
“Is this a repeated pattern, or one trade in a small sample?”
Start with what changed.
Get a briefing on the tokens and wallets you follow. Each item includes what happened, why it matched your interests and the records that support it.
“What changed overnight, and which findings need another look?”
Ask what you could have known then.
Replay a saved rule using information available at the selected time. Show data-availability gaps so later labels and discoveries do not silently improve the result.
“Would this rule have triggered before the event, using only records available then?”
Share the investigation with its sources.
Keep team watchlists, private notes and labels together. Export a dated bundle of findings, source references, research settings and coverage for another analyst or agent to inspect.
“Give my team the findings and the evidence needed to review them.”
Know the limits
More depth. A defined budget.
Current service ceilings are shown below. Holdings thresholds and the token-to-credit rate are configured by the service. A larger history window does not guarantee complete historical data.
| Allowance | Free | Holder | Pro | Deep job |
|---|---|---|---|---|
| Watched addresses | 5 | 25 | 100 | Current tier |
| Concurrent jobs | 1 | 2 | 4 | Current tier |
| RPC calls per job | 25 | 100 | 300 | 1,000 |
| Observation rows per job | 25 | 100 | 300 | 1,000 |
| Maximum graph depth | 1 | 3 | 4 | 6 |
| Maximum requested history | 7 days | 31 days | 90 days | 180 days |
| Credits at admission | 0 | 0 | 0 | 10 |
Shared service and provider limits also apply, including a deployment-wide cap of 100 distinct watched addresses. Deep jobs pay for admitted work, not a guaranteed finding. Cancellation, expiry and incomplete results do not refund an admitted job. A job cannot outlive its authorization grant.
What evidence means
Know what a result proves.
Does a funding connection prove common ownership?
No. A transfer or shared funding source is evidence of activity. It does not establish that the same person controls both wallets. Exchange withdrawals and shared services can connect unrelated users.
Are evidence receipts onchain payment proofs?
No. Our receipts store research results within the service. They are not independently signed settlement proofs. Source transaction references help you inspect the underlying activity.
Can I buy any external agent tool here?
This is a specialised intelligence service, not a marketplace for third-party tools. Agents access its research through MCP and scoped HTTP tools.
Are the planned workflows available now?
No. Wallet-group signals, persistent research objectives, dossiers, launch timelines, reputation histories, briefings, historical replay and team evidence bundles require further backend work.