MCP server for Bitpanda — portfolio, wallets, trading, market data
bitpanda-mcp is a solid financial API wrapper with good naming conventions and HTTP transport. All 7 tools follow consistent verb_noun patterns (get_*, list_*) and are READ_ONLY, which is appropriate. Tool descriptions are present but generic, they state WHAT without explaining WHEN to use each tool or how they differ. Parameter schemas are mostly visible in the fastmcp registration (via Python type hints in function signatures), but descriptions for individual parameters are sparse or missing entirely. Output schemas are not formally documented. Error handling is basic. The server properly uses tool annotations (readOnlyHint=True) and stateless HTTP, showing current spec alignment.
Get asset metadata
Get portfolio overview with EUR values
Check a specific asset price by symbol
List held-asset ticker prices or a capped market-wide ticker list
List normalized buy/sell trading activity
List asset transactions including deposits, withdrawals, transfers, and trades
List all asset wallets with balances
Parameter descriptions are missing or minimal. Tools like get_price, list_prices, and list_trades accept parameters (symbol, limit, offset) but the parameter descriptions are not visible in server.py or the tool module docstrings provided. LLMs cannot infer parameter meaning from names alone, 'symbol' could mean ticker symbol, internal symbol, or display symbol.
Output schemas are not documented. Tool descriptions state WHAT they return (e.g. 'List held-asset ticker prices') but do not specify the structure of the response, field names, types, whether results are paginated, what pagination looks like, or whether a count is included. Without this, LLMs cannot plan downstream tool calls or extract data reliably.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 51 | 2026-07-28+ | v2 |
| 2026-03-09 | B | 71 | - | v1 |
Tool descriptions lack actionable context about WHEN to use each tool and how they differ. 'Get portfolio overview with EUR values' and 'List all asset wallets with balances' sound similar, the LLM may not know when to call get_portfolio vs list_wallets. Descriptions should explain: get_portfolio aggregates ALL holdings with EUR totals; list_wallets shows per-wallet balances.
No pagination guidance for list_* tools. 'List held-asset ticker prices or a capped market-wide ticker list' mentions a cap but does not say what the cap is, whether there is pagination, or how to retrieve more results. If the API supports limit/offset or cursor-based pagination, the tool description and parameter descriptions must explain it.
Error handling does not guide LLM recovery. No guidance is visible for what happens when an API key is invalid, rate-limited, or when a symbol/wallet does not exist. LLMs need actionable errors: 'Symbol not found. Available symbols: BTC, ETH, ...' rather than a raw error code.
No explicit parameter constraints (enums, ranges, patterns) visible for key parameters. For example, 'symbol' parameter: is it case-sensitive? Does it accept 'BTC' and 'bitcoin' or only ticker symbols? Are there allowed values? This forces LLMs to guess and can cause invalid API calls.