Docs

CaudexCatena MCP Documentation

Connect CaudexCatena MCP to your AI client and run your first successful prompt.

Setup

The CaudexCatena MCP server URL is the same for every client:

MCP server URL

https://mcp-api.caudexcatena.com/mcp

Open Connectors. In Claude, go to Settings → Connectors → Add custom connector.
Add the CaudexCatena MCP server. Paste the CaudexCatena MCP URL below as the connector URL and save.
MCP server URL
```
https://mcp-api.caudexcatena.com/mcp
```
Authorize CaudexCatena. Use the connector authorization flow and sign in with your WorkOS account.
Complete the WorkOS authorization prompt in your client.

Go to Apps. Open Settings → Apps → Advanced settings.
Create App. Provide the CaudexCatena MCP URL below as the server endpoint.
MCP server URL
```
https://mcp-api.caudexcatena.com/mcp
```
Authorize CaudexCatena. Use the app authorization flow and sign in with your WorkOS account.
Complete the WorkOS authorization prompt in your client.

Open MCP settings. In Cursor, go to Settings → Tools & MCP (or edit ~/.cursor/mcp.json directly).

Add the CaudexCatena MCP server. Paste the snippet below into your MCP configuration, then save.

~/.cursor/mcp.json

{
  "mcpServers": {
    "caudexcatena_mcp": {
      "url": "https://mcp-api.caudexcatena.com/mcp"
    }
  }
}

mcp.json showing the CaudexCatena MCP server entry

Authorize CaudexCatena. Use the client authorization flow and sign in with your WorkOS account.
Complete the WorkOS authorization prompt in your client.

Open your client's MCP settings. Most MCP-capable clients accept a remote server URL and OAuth authorization.

Add the CaudexCatena MCP server. Use the CaudexCatena MCP URL below as the remote server endpoint or add the snippet below to the config file.

MCP server URL

https://mcp-api.caudexcatena.com/mcp

mcp config

{
  "mcpServers": {
    "caudexcatena_mcp": {
      "url": "https://mcp-api.caudexcatena.com/mcp"
    }
  }
}

Authorize CaudexCatena. Use the client's OAuth authorization flow and sign in with your WorkOS account.
Complete the WorkOS authorization prompt in your client.

Ready to use

Ask investigation questions and review source context in your AI client. Try a first prompt like this one:

Example

Using CaudexCatena MCP, summarize BTC address bc1q…7k9x: role, exposure, and suggested next checks.

Docs

Troubleshooting

Common setup issues and how to resolve them.

CaudexCatena MCP doesn't appear in my client

Confirm the server URL is exact and your client supports remote MCP servers with OAuth authorization. Fully quit and reopen the client so it reloads its MCP configuration.

Authorization failed

Reconnect the MCP server and complete the WorkOS authorization prompt with the same account that owns the subscription.

Calls return no data

A successful call can still return no matches for an address or transaction. Check the identifier and chain. Failed calls, retries, and tool discovery are never charged.

I've used my included calls

Upgrade to a higher plan for immediate effect, or add extra-call balance when included calls are unavailable. Included calls are always used before extra-call balance.

Docs

Investigation Examples

Review practical prompt patterns, expected response structure, and follow-up flows for CaudexCatena MCP.

Your AI client may also use authorized case documents, investigation notes, screenshots, web search results, and other sources you provide. CaudexCatena MCP supplies blockchain intelligence; the AI client can combine that context with other material to produce a richer draft for human review.

Address Summary

Source inputs. One BTC address. Optionally, case notes you are authorized to use.

Example prompt

Prompt

Can you profile BTC address bc1qxy2kgdygjrsqtzq2n0yrf2493p83kkfjhx0wlh? Summarize activity, incoming and outgoing exposure, notable counterparties, and the next records worth reviewing.

Expected response pattern

Address profile with label, asset, first activity, last observed activity, received amount, spent amount, and remaining balance.
Activity summary that separates meaningful historical movement from recent dust-like activity.
Incoming and outgoing exposure summaries.
Notable counterparties or labels.
Next records worth reviewing.
A guardrail that signals are context, not conclusions.

Sample response pattern

An illustrative pattern, not a live query. Actual output depends on the records available and your AI client. Treat it as a draft for human review.

What CaudexCatena MCP may return:

Address profile: display label, primary asset, first and last observed activity, total received, total spent, and remaining balance.
Exposure by counterparty category, split into incoming and outgoing direction.
A counterparty and label list for the most significant related addresses.
The underlying transaction records behind each figure.

How the AI client may summarize it:

A profile block covering the label, asset, activity dates, and balances.
An activity summary that separates meaningful historical movement from recent dust-like activity.
Incoming and outgoing exposure described in plain language.
Notable counterparties or labels worth flagging.
A short list of the next records worth reviewing.
A guardrail line noting that signals are context, not conclusions.

What to review before action:

Open the raw records behind any figure before relying on it.
Treat dust-like or low-value activity as possible noise rather than intent.
Labels and categories are signals — they do not prove ownership, control, or wrongdoing.

Possible follow-up prompts

Break down exposure for one specific counterparty in more detail.
List the largest outputs from this address and where they moved next.
Check whether any counterparty is a known swap or bridge service.

Cross-Chain Check

Source inputs. One Ethereum transaction hash. Optionally, notes on the suspected bridge or swap.

Example prompt

Prompt

I'm reviewing Ethereum transaction 0xa7f50be6aa7de66a66d4b646205ff1f936ad5936c0a50057695618c75aebcc56. Check whether CaudexCatena has bridge or swap context for it, then summarize both sides of the movement and what I should verify next.

Expected response pattern

A short answer that CaudexCatena has swap or bridge context.
Identification of the service or protocol, such as ChainFlip, where available.
Ethereum-side settlement summary.
BTC-side source movement summary.
A recipient table for the Ethereum recipients.
An optional swap path map.
A verification list for open items and caveats.

Sample response pattern

An illustrative pattern, not a live query. Actual output depends on the records available and your AI client. Treat it as a draft for human review.

What CaudexCatena MCP may return:

The Ethereum transaction record with inputs, outputs, and fees.
Swap or bridge context for the transaction, where CaudexCatena has it.
Identification of the service or protocol involved, such as ChainFlip, where available.
Settlement detail on the Ethereum side and source movement on the BTC side.
The set of Ethereum recipients for the movement.

How the AI client may summarize it:

A short answer on whether CaudexCatena has swap or bridge context.
The service or protocol name, where identified.
An Ethereum-side settlement summary.
A BTC-side source movement summary.
A recipient table for the Ethereum recipients.
An optional swap path map across the two sides.
A verification list of open items and caveats.

What to review before action:

Confirm both legs describe the same movement before linking them.
Service identification is a signal; verify against the service's own records where possible.
Check that timing and amounts on each side reconcile.

Possible follow-up prompts

Trace the BTC-side source further back for origin context.
Profile the largest Ethereum recipient from this movement.
Confirm the swap path with the matching transaction on the other side.

Hypothesis Check

Source inputs. One Bitcoin transaction hash and your working hypothesis to test.

Example prompt

Prompt

Test this hypothesis for Bitcoin transaction a44c105dba43459f9a526d9263157ab1e10835afeb5940133411a4a571a42ce5: it may be a split point before funds continue along more than one path. Separate observed facts, weak signals, and what would disprove the hypothesis.

Expected response pattern

Assessment of whether the hypothesis is supported, weakened, or not supported.
Observed facts about the transaction structure.
A compact table for outputs and next observed movement.
Weak signals and interpretation.
What would disprove or weaken the hypothesis.
A bottom-line wording that is defensible and uncertainty-aware.

Sample response pattern

An illustrative pattern, not a live query. Actual output depends on the records available and your AI client. Treat it as a draft for human review.

What CaudexCatena MCP may return:

The Bitcoin transaction structure: inputs, outputs, and amounts.
The next observed movement for each output.
Exposure and label context for the outputs, where available.

How the AI client may summarize it:

An assessment of whether the hypothesis is supported, weakened, or not supported.
Observed facts about the transaction structure.
A compact table for outputs and their next observed movement.
Weak signals and how to interpret them.
What would disprove or weaken the hypothesis.
A defensible, uncertainty-aware bottom line.

What to review before action:

Keep observed facts separate from interpretation.
A split pattern is common and is not proof of a specific intent.
Follow downstream movement before drawing a conclusion.

Possible follow-up prompts

Trace each output path forward one more hop.
Test an alternative reading, such as consolidation or a change output.
Compare the structure against known service or wallet patterns.

Only provide documents or case context you are authorized to use. Third-party AI clients, web search tools, and local files are governed by their own settings and policies. CaudexCatena MCP calls consume included calls only when CaudexCatena data is successfully returned. Local file parsing, web research, and prose generated from already retrieved context do not consume CaudexCatena MCP calls.