SimianX AI

Start deep research

start_deep_analysis

Start the deep research committee on a stock or a coin: four analysts (fundamentals, market, news, and positioning for stocks or derivatives for crypto), a bull and a bear researcher, a research manager and a risk manager, ending in a verdict with entry, stop and target and a PDF. Takes about ten minutes; each analyst, each researcher, the start and end of every debate round, the verdict and the PDF link arrive in get_updates (or the run's events) as they land. After starting, follow it until done and tell the user every finished step; never wait for the user to ask. (uses points, metered on the models used like the website)

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "symbol": "<symbol>"
    },
    "name": "start_deep_analysis"
  }
}

Input Schema

Name	Required	Description
`asset`	No	stock (default) or crypto
`agents`	No	Subset of the committee seats (default all)
`models`	No	Only when the user asks for specific models: agent -> model id from list_models (paid plans). Leave out to run on the user's saved setup (get_settings)
`symbol`	Yes	Stock ticker (NVDA) or coin (BTC); crypto also takes a pair id such as BTC-USDT-SWAP
`horizon`	No	Holding horizon (default swing)
`language`	No	Language of the write-ups, e.g. en, zh-CN, ja (default: the account's)
`idempotency_key`	No	Optional: any unique string for this action; send the same one when retrying the same call and the first result comes back instead of the action running twice

Schema Changelog

Changes observed during successful MCP inspections.

Changed3 schema fields changedNov 13, 2026

addedInput schema / properties / idempotency_key

Added value: +{
+  "description": "Optional: any unique string for this action; send the same one when retrying the same call and the first result comes back instead of the action running twice",
+  "maxLength": 128,
+  "type": "string"
+}

changedInput schema / properties / models / description

Previous value: -"Per-agent model override, agent -> model id from list_models (paid plans)"New value: +"Only when the user asks for specific models: agent -> model id from list_models (paid plans). Leave out to run on the user's saved setup (get_settings)"

removedInput schema / properties / tier

Removed value: -{
-  "description": "Model tier (default fast)",
-  "enum": [
-    "fast",
-    "standard",
-    "frontier"
-  ],
-  "type": "string"
-}

First observedOct 16, 2026

TDQS

A4.1/5.0

Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations cover only the generic safety profile (readOnlyHint=false, openWorldHint=true, destructiveHint=false), and the description goes well beyond them: a ~10 minute duration, asynchronous step-by-step emission of analysts/researchers/debate rounds/verdict/PDF via get_updates, and metered point consumption. It does not say what happens on failure or how to resume/interrupt a run, but the operational shape is clear.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loads what the tool does and the deliverable, then the timing and follow-up obligation. Dense and mostly waste-free, though the trailing parenthetical about points and the long committee enumeration make it slightly sprawling for a single paragraph.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a long-running, multi-agent, 7-parameter tool with no output schema, the description supplies the missing pieces: expected runtime, where results surface, and the obligation to poll and report. Minor gaps remain on failure/abort behavior and what the final verdict payload looks like beyond the PDF link.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100% and the descriptions are already rich (asset enum, agents subset, models mapping, symbol formats, horizon, language, idempotency_key retry semantics). The description adds essentially nothing beyond the schema's own text, so baseline 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ('Start the deep research committee on a stock or a coin') and enumerates the committee composition and the deliverable (verdict with entry, stop, target, plus a PDF). It does not explicitly contrast itself with nearby siblings such as run_live_analysis or start_autopilot, so sibling differentiation is only partial (via the get_updates follow-up pointer), keeping it just shy of a 5.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit when-to-use and how-to-follow guidance: 'After starting, follow it until done and tell the user every finished step; never wait for the user to ask,' with get_updates named as the polling channel. It also states a when-not for the models parameter ('Only when the user asks for specific models') and warns about point costs, so the agent knows both the workflow and the constraint.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All SimianX AI tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.