Connect AI agents to the Nimble MCP server

Vendor MCP

The Nimble MCP connector routes your AI agent's tool calls to Nimble's own MCP server through Scalekit. Each user connects their own Nimble access token once, and Scalekit sends it with every call, so your agent never handles credentials. It comes with 27 tools.

Tools: 27
What they doRead · write · destructive: 16 · 6 · 516 read6 write5 destructive
Users sign in with: Access token

Setup

Install the SDK
- Node.js
- Python
Terminal window
npm install @scalekit-sdk/node dotenv
Terminal window
pip install scalekit-sdk-python python-dotenv
Set your credentials
Add your Scalekit credentials to your .env file. Find values in app.scalekit.com > Developers > API Credentials.
.env
```
SCALEKIT_ENVIRONMENT_URL=<your-environment-url>
SCALEKIT_CLIENT_ID=<your-client-id>
SCALEKIT_CLIENT_SECRET=<your-client-secret>
```
Create the Nimble MCP connection
In AgentKit > Connections, create a Nimble MCP connection. The name you give it is the connection_name your code passes. See Configure connections.
Console steps with screenshots
Generate an API key from your Nimble account so Scalekit can authenticate your users’ web intelligence requests.
1. Create an API key in Nimble
  - Sign in to app.nimbleway.com.
  - In the left sidebar, click API Playground or look for the Create API Key item in the onboarding checklist at the bottom left.
  - Click Create API Key.
  In the Create API Key dialog:
  - Enter a Key name that identifies this integration, for example Agent Auth.
  - Click Create key.
  Nimble displays your new API key. Copy it immediately — it is only shown once.
  
  Nimble shows the full API key only on the creation screen. If you navigate away without copying it, you must revoke the key and generate a new one.
2. Add the API key to Scalekit
  - In the Scalekit dashboard, go to AgentKit > Connections > Create Connection.
  - Search for Nimble MCP and click Create.
  - In the API Key field, paste the key you copied.
  - Click Save.
  Scalekit validates the key and marks the connection as active. Your users can now execute Nimble tools through their own connected accounts.

Authorize a user and make your first call

Node.js
Python

1
import { ScalekitClient } from '@scalekit-sdk/node'
2
import 'dotenv/config'
3
import { createInterface } from 'node:readline/promises'
4

5
const scalekit = new ScalekitClient(
6
  process.env.SCALEKIT_ENVIRONMENT_URL,
7
  process.env.SCALEKIT_CLIENT_ID,
8
  process.env.SCALEKIT_CLIENT_SECRET,
9
)
10
const actions = scalekit.actions
11

12
const connector = 'nimblemcp'
13
const identifier = 'user_123'
14

15
// Generate an authorization link for the user
16
const { link } = await actions.getAuthorizationLink({ connectionName: connector, identifier })
17
console.log('Authorize Nimble MCP:', link)
18
const rl = createInterface({ input: process.stdin, output: process.stdout })
19
await rl.question('Press Enter after authorizing...')
20
rl.close()
21

22
// Make your first call
23
const result = await actions.executeTool({
24
  connector,
25
  identifier,
26
  toolName: 'nimblemcp_nimble_agent_templates_list',
27
  toolInput: {},
28
})
29
console.log(result)

npx tsx quickstart.mts

1
import os
2
from scalekit import ScalekitClient
3
from dotenv import load_dotenv
4
load_dotenv()
5

6
scalekit_client = ScalekitClient(
7
    env_url=os.getenv("SCALEKIT_ENVIRONMENT_URL"),
8
    client_id=os.getenv("SCALEKIT_CLIENT_ID"),
9
    client_secret=os.getenv("SCALEKIT_CLIENT_SECRET"),
10
)
11
actions = scalekit_client.actions
12

13
connection_name = "nimblemcp"
14
identifier = "user_123"
15

16
# Generate an authorization link for the user
17
link_response = actions.get_authorization_link(
18
    connection_name=connection_name,
19
    identifier=identifier,
20
)
21
print("Authorize Nimble MCP:", link_response.link)
22
input("Press Enter after authorizing...")
23

24
# Make your first call
25
result = actions.execute_tool(
26
    tool_input={},
27
    tool_name="nimblemcp_nimble_agent_templates_list",
28
    connection_name=connection_name,
29
    identifier=identifier,
30
)
31
print(result)

python quickstart.py

Each user opens the link once and enters their Nimble MCP server credentials there. If your app already has a user's credentials, add the account from your backend instead, as the console steps above show. See Authorize a user for the full flow and statuses.

Tools

Pass the exact name to execute_tool

Try in Playground Request a tool

nimblemcp_nimble_agent_templates_getGet full details of a pre-built agent template.Read-only
Agent Templates Get
Get full details of a pre-built agent template. When to use: - You have a template name from `nimblemcp_nimble_agent_templates_list` and want its full definition before basing a new agent on it. When NOT to use: - You don't know the template name — use `nimblemcp_nimble_agent_templates_list` first.
Inputs
template_namestringrequired
Template name as returned by nimble_agent_templates_list.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_agent_templates_listList pre-built agent templates that can seed `nimblemcp_nimble_agents_create` via its `template` field.Read-only
Agent Templates List
List pre-built agent templates that can seed `nimblemcp_nimble_agents_create` via its `template` field. When to use: - You want to see what pre-built agent templates exist before creating a custom agent. When NOT to use: - You already have a template name — use `nimblemcp_nimble_agent_templates_get` to inspect it.
Inputs
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
limitinteger
Maximum number of templates to return (1-100).default 50
offsetinteger
Offset for pagination.default 0
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_agents_getGet full details of a specific agent including its input/output schema.Read-only
Agents Get
Get full details of a specific agent including its input/output schema. Use after `nimblemcp_nimble_agents_list` to inspect an agent before running it. When to use: - You have an agent name and need to see its full schema before running it. When NOT to use: - You don't know the agent name — use `nimblemcp_nimble_agents_list` first.
Inputs
agent_idstringrequired
Agent id as returned by nimble_agents_list or nimble_agents_create.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_agents_listBrowse the catalog of pre-built Nimble agents.Read-only
Agents List
Browse the catalog of pre-built Nimble agents. Use this as the first step when you want to run a structured extraction on a known site or data source. When to use: - You want to discover available agents for a specific domain or data source. - You want to paginate through all agents in the catalog.
Inputs
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
limitinteger
Maximum number of agents to return in one page (1-100).default 50
offsetinteger
Offset for pagination. Use 0 for the first page.default 0
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_agents_run_resultFetch the output of a completed agent run.Read-only
Agents Run Result
Fetch the output of a completed agent run. Returns the run's output (text or structured JSON with trust/citation metadata). Call after `nimblemcp_nimble_agents_run_status` reports a terminal status. When to use: - An agent run has completed and you need its final output. When NOT to use: - The run may still be in progress — check `nimblemcp_nimble_agents_run_status` first.
Inputs
agent_idstringrequired
Agent id the run belongs to.
run_idstringrequired
Run id returned by nimble_agents_run.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_agents_run_statusCheck the status of an agent run.Read-only
Agents Run Status
Check the status of an agent run. While the status is not terminal (e.g. still pending/running), poll again after ~15-30 seconds. Once completed, fetch the output via `nimblemcp_nimble_agents_run_result`. When to use: - You started a run with `nimblemcp_nimble_agents_run` and want to check whether it has finished. When NOT to use: - You want crawl status — use `nimblemcp_nimble_crawl_status`. - You want extract template generation status — use `nimblemcp_nimble_extract_templates_generation_status`.
Inputs
agent_idstringrequired
Agent id the run belongs to.
run_idstringrequired
Run id returned by nimble_agents_run.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_agents_runs_listList past and in-progress runs of an agent.Read-only
Agents Runs List
List past and in-progress runs of an agent. When to use: - You want to see run history for a specific agent, including in-progress runs. When NOT to use: - You want details of one specific run — use `nimblemcp_nimble_agents_run_status` or `nimblemcp_nimble_agents_run_result`.
Inputs
agent_idstringrequired
Agent id whose runs to list.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
limitinteger
Maximum number of runs to return (1-100).default 50
offsetinteger
Offset for pagination.default 0
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_crawl_listList crawl jobs, optionally filtered by status.Read-only
Crawl List
List crawl jobs, optionally filtered by status. When to use: - You want to see all active or past crawl jobs in the account. - You want to filter crawls by their current status.
Inputs
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
limitinteger
Maximum number of crawl jobs to return (default: 20).default 20
statusstring
Optional filter by crawl status: - "queued": Waiting to start - "running": Currently in progress - "succeeded": Completed successfully - "failed": Completed with errors - "canceled": Manually terminated - "all": Show all statuses (default if not specified)
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_crawl_statusCheck the status and progress of a running or completed crawl job.Read-only
Crawl Status
Check the status and progress of a running or completed crawl job. When to use: - You started a crawl with `nimblemcp_nimble_crawl_run` and want to check its progress or retrieve results. When NOT to use: - You want async task results from `nimble_extract_async` or `nimble_agent_run_async` — use `nimblemcp_nimble_task_results` for those.
Inputs
crawl_idstringrequired
The unique crawl identifier returned by nimble_crawl_run.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
Optional description of the tool's purpose.
tool_reasoningstring
Optional reasoning about why this tool was selected.
nimblemcp_nimble_extractExtract and parse content from a specific URL using Nimble's Extract API.Read-only
Extract
Extract and parse content from a specific URL using Nimble's Extract API. This is a synchronous call — it waits for extraction to complete before returning. When to use: - You have a specific URL and need its content immediately. - You want structured content (markdown, text) from a single page. When NOT to use: - You need to extract many pages — use `nimblemcp_nimble_crawl_run` instead. - The page loads slowly — use `nimblemcp_nimble_extract_async` for async extraction.
Inputs
urlstringrequired
The URL to extract content from. Must be a valid, accessible web page.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
countrystring
Country code for geo-targeted results (e.g., "US", "GB", "IL"). Defaults to "US".
driverstring
Optional browser driver tier for web extraction: - "vx6": fast HTTP requests, no JavaScript â static pages and APIs - "vx8": headless browser with JavaScript rendering â dynamic pages and SPAs - "vx8-pro": headful browser with JavaScript rendering â complex interactions - "vx10": stealth headless browser â sites with strict bot mitigation - "vx10-pro": stealth headful browser â heavily bot-protected sites If omitted, the backend selects. Escalate explicitly for JavaScript-heavy pages ("vx8"/"vx8-pro"), or for sites with strict bot mitigation ("vx10"/"vx10-pro"), when a default extraction returns empty or truncated content.
localestring
Language/locale code (e.g., "en", "fr", "de"). Defaults to "en".
output_formatstring
Content output format for the extracted page: - "markdown": Markdown-formatted content with structure (default) - "plain_text": Plain text without formatting - "simplified_html": Clean HTML without scripts/stylesdefault markdown
tool_descriptionstring
No description.
tool_reasoningstring
No description.
waitstring
Optional wait time in milliseconds before extracting content. Useful for JavaScript-rendered pages that need time to load. Example: wait=3000 for pages that render content dynamically.
nimblemcp_nimble_extract_templates_generation_statusCheck the current status of an extract template generation.Read-only
Extract Templates Generation Status
Check the current status of an extract template generation. Status flow: in-progress (pending/processing) → completed or failed. On completion, `template_name` and `version_id` are set and you can proceed to run the template. On failure, inspect `error` for diagnostics. When to use: - You started a generation with `nimblemcp_nimble_extract_templates_generate` or `nimblemcp_nimble_extract_templates_update_from_template` and want to check if it completed. When NOT to use: - You want crawl job status — use `nimblemcp_nimble_crawl_status`. - You want async task results — use `nimblemcp_nimble_task_results`.
Inputs
generation_idstringrequired
Server-assigned id returned by nimble_extract_templates_generate or nimble_extract_templates_update_from_template.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_extract_templates_getGet full details of a specific extract template including its input/output schema.Read-only
Extract Templates Get
Get full details of a specific extract template including its input/output schema. Use after `nimblemcp_nimble_extract_templates_list` to inspect a template before running it. The response includes the template's input_schema / output_schema, description, and metadata. When to use: - You have a template name and need to see its full schema before running it. When NOT to use: - You don't know the template name — use `nimblemcp_nimble_extract_templates_list` first.
Inputs
template_namestringrequired
Exact template name as returned by nimble_extract_templates_list.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_extract_templates_listBrowse the catalog of extract templates.Read-only
Extract Templates List
Browse the catalog of extract templates. Use this tool as the first step to find an existing template for a data collection task. Each template is purpose-built for a specific website or data type (e.g. Amazon products, LinkedIn profiles, Google Maps reviews). Typical workflow: list → get (inspect details) → run (collect data). When to use: - You want to discover available extract templates for a specific domain or data source. - You want to paginate through all templates in the catalog.
Inputs
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
limitinteger
Maximum number of templates to return in one page (1-100).default 50
querystring
Short keyword to filter templates by name or domain (e.g. 'amazon', 'linkedin'). Use 1-2 words for best results.
skipinteger
Offset for pagination. Use 0 for the first page, then pass the previous response's next_skip â unrunnable templates are skipped over, so skip + limit is not the next offset.default 0
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_mapDiscover all URLs on a website by crawling its pages and sitemap.Read-only
Map
Discover all URLs on a website by crawling its pages and sitemap. Returns a flat list of URLs found on the site. Useful for understanding site structure before targeted extraction. When to use: - You need to enumerate the URL space of a site before deciding what to extract. - You want a quick snapshot of all discoverable links on a domain. When NOT to use: - You need the page content itself — use `nimblemcp_nimble_crawl_run` or `nimblemcp_nimble_extract`.
Inputs
urlstringrequired
The website URL to map. The tool will discover all linked pages starting from this URL.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
domain_filterstring
Control URL scope for discovery: - "domain": Only URLs on the exact domain (default) - "subdomain": Include URLs from subdomains - "all": Include all discovered URLs regardless of domain
limitstring
Maximum number of URLs to return. If not specified, the API default is used.
sitemapstring
How to use the site's sitemap for URL discovery: - "include": Use sitemap in addition to link following (default) - "skip": Ignore sitemap, only follow links - "only": Only return URLs found in sitemap
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_searchSearch the web using Nimble's Search API with configurable content richness.Read-only
Search
Search the web using Nimble's Search API with configurable content richness. When to use: - You need web search results with optional AI-generated answers. - You want relevance-ranked results with snippet or full-page content. When NOT to use: - You need to extract the full content of a known URL — use `nimblemcp_nimble_extract`.
Inputs
querystringrequired
The search query to execute. Can be natural language questions or keywords.
content_typestring
Filter by content type (only supported with focus="general"). Supports semantic groups ("documents", "spreadsheets", "presentations") and specific formats ("pdf", "docx", "xlsx", etc.).
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
countrystring
Country code for geo-targeted results (e.g., "US", "GB", "IL"). Defaults to "US".
end_datestring
Filter results before this date. Format: YYYY-MM-DD or YYYY. Cannot be combined with time_range.
exclude_domainsstring
Blacklist of domains to exclude from search results (max 50). Results from these domains will be filtered out. Example: ["pinterest.com", "facebook.com"]
focusstring
Search focus mode that determines content sources and ranking. Can be a single mode string or a list of explicit Extract Template names. Single mode options: - "general": Standard web search (default) - "news": Real-time news search - "location": Location-based search - "shopping": E-commerce and product search (AI-powered) - "geo": Generative engine optimization (AI-powered) - "social": Social media content search (AI-powered) Or a list of explicit Extract Template names (e.g., ["amazon_serp", "target_serp"]) to skip the AI router and query exactly those templates. Only the templates behind the shopping, social and geo modes are accepted. Max 10 items, no duplicates.default general
full_contentboolean
Return full page content per result instead of snippets only. How that content is fetched depends on the depth it is paired with: with search_depth="lite" each result is scraped and parsed in full; with search_depth="standard" it instead widens live crawling across web and news sources so results carry full markdown. Either way it costs more and is slower (5-15 seconds) â use it to read deeply into that same closed lookup, not as a substitute for a Web Search Agent.default false
include_answerboolean
Nimble Answer â one synthesized, citation-backed response for a closed factual question. Default is False (raw Search results). Still not a substitute for a Web Search Agent on research or datasets.default false
include_domainsstring
Whitelist of domains to include in search results (max 50). Only results from these domains will be returned. Example: ["wikipedia.org", "britannica.com"]
localestring
Language/locale code (e.g., "en", "fr", "de"). Defaults to "en".
max_resultsinteger
Maximum number of search results to return (1-100). Default is 3. Actual count may be less depending on available results. - Use 3-5 for a closed lookup - Use up to 10 only when that single search still answers the ask - Do not raise this to "cover more ground" â that fan-out is a Web Search Agentdefault 3
max_subagentsinteger
Maximum number of Extract Templates the AI router picks and runs in parallel for the shopping, social and geo focus modes (1-10). Default is 3. Ignored for an explicit template list (every listed template runs) and for general, news, location modes.default 3
output_formatstring
Content output format for extracted pages: - "markdown": Markdown-formatted content with structure (default) - "plain_text": Plain text without formatting - "simplified_html": Clean HTML without scripts/stylesdefault markdown
search_depthstring
Content richness level. Omit to let the Nimble API choose the default. - "lite": Metadata only (title, description, URL). Fast (1-3 seconds). Use for quick lookups, getting links, or simple fact-checking. - "standard": Rich content snippets (~2K chars per result) optimized for AI agents. Only supported with focus="general" â every other focus mode is rejected with a 422, so pair those with "lite" instead.
start_datestring
Filter results after this date. Format: YYYY-MM-DD or YYYY. Cannot be combined with time_range.
time_rangestring
Filter by recency with predefined periods. Cannot be combined with start_date/end_date. Options: "hour", "day", "week", "month", "year"
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_task_resultsGet the status and results of an async task.Read-only
Task Results
Get the status and results of an async task. Use this to retrieve results from `nimblemcp_nimble_extract_async` or `nimblemcp_nimble_agent_run_async` after initiating them. When to use: - You started an async extraction or agent run and want to poll for its results. When NOT to use: - You want crawl status — use `nimblemcp_nimble_crawl_status` for crawl jobs.
Inputs
task_idstringrequired
The task ID returned by nimble_extract_async or nimble_extract_templates_run_async.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_agents_createCreate a new web search agent.Write
Agents Create
Create a new web search agent. All parameters are optional; pass `template` to start from a pre-built agent template (see `nimblemcp_nimble_agent_templates_list`), or describe the agent via name/goals/sources. When to use: - You want a persistent, reusable agent configured with specific goals, sources, and output schema. When NOT to use: - You just need a one-off custom extraction from a prompt and URL — use `nimblemcp_nimble_extract_templates_generate` instead.
Inputs
agent_namestring
Unique name for the agent.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
descriptionstring
What the agent does.
display_namestring
Human-friendly display name.
effortstring
Set from the task â do not ask the user to pick: 'low'/'medium' for fast simple asks; 'high' default once several fields need digging; 'x-high'/'max' for complex profiles (financials, filings, history). dataset_building requires high or above. Optional preview at low, then re-run at the recommended tier.
output_schemastring
JSON Schema for the agent's structured output, stored and reused by every run. Required when use_case is 'dataset_building' or 'enrichment'. Must be real JSON Schema (Draft 7), not a field->example map: each field is a subschema like {"type": "string"}. Root is "object" with non-empty "properties", or "array" whose "items" is an object (a table: prefer it for lists, dataset_building, and list input_data). Put "required" next to "properties", never inside it. Max nesting depth 5 (objects and arrays both count): prefer flat columns over nested arrays. Max 100 properties total; no format/pattern/min*/max* keywords; no standalone "null" type (use ["string", "null"]). Example: {"type": "array", "items": {"type": "object", "properties": {"name": {"type": "string"}}, "required": ["name"]}}.
skillstring
Skill/domain the agent specializes in.
sourcesstring
Source preferences: {allow: [...], block: [...], prioritize: str, avoid: str}.
suggested_questionsstring
Example questions for the agent.
templatestring
Optional Web Search Agent gallery template name from nimble_agent_templates_list. Not an Extract Template name.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
use_casestring
One of: 'research' (dossiers, briefings, synthesis), 'dataset_building' (entity lists, catalogs, extract-all without a URL), 'enrichment' (add fields to records the user already has). Classify the kind first.
nimblemcp_nimble_agents_runStart an asynchronous agent run.Write
Agents Run
Start an asynchronous agent run. Returns immediately with a run `id` and `status`. Poll `nimblemcp_nimble_agents_run_status` until the status is terminal, then fetch the output via `nimblemcp_nimble_agents_run_result`. When to use: - You want to run a pre-built or custom agent (existing or create-or-reuse by name) against a research question. When NOT to use: - You want a synchronous, one-off custom extraction from a URL and prompt — use `nimblemcp_nimble_extract_templates_run` instead.
Inputs
inputstringrequired
Natural-language research question or task for the agent.
agent_idstring
Existing agent id to run. Provide this or `agent_name`, not both.
agent_namestring
Stable agent name for create-or-reuse: an unseen name creates a new agent, an existing name reuses it and returns the same agent id across calls. Ignored if `agent_id` is set.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
effortstring
Run override. Set from the task â do not ask the user to pick: 'low'/'medium' for fast simple asks; 'high' default once several fields need digging; 'x-high'/'max' for complex profiles (financials, filings, history). dataset_building requires high or above. Optional preview at low, then re-run at the recommended tier.
input_datastring
Optional structured input for enrichment runs: up to 40 rows, with rows multiplied by new output columns capped at 200 cells.
output_schemastring
Usually omit: the run uses the agent's stored schema. Set the schema once on the agent rather than regenerating it per run. Required when this call creates a 'dataset_building' or 'enrichment' agent via `agent_name`. Must be real JSON Schema (Draft 7), not a field->example map: each field is a subschema like {"type": "string"}. Root is "object" with non-empty "properties", or "array" whose "items" is an object (a table: prefer it for lists, dataset_building, and list input_data). Put "required" next to "properties", never inside it. Max nesting depth 5 (objects and arrays both count): prefer flat columns over nested arrays. Max 100 properties total; no format/pattern/min*/max* keywords; no standalone "null" type (use ["string", "null"]). Example: {"type": "array", "items": {"type": "object", "properties": {"name": {"type": "string"}}, "required": ["name"]}}.
skillstring
Skill override for this run. One-time only against an existing agent; becomes the new agent's stored skill if this call creates it via `agent_name`.
sourcesstring
Source preferences override: {allow: [...], block: [...], prioritize: str, avoid: str}.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
use_casestring
One of: 'research' (dossiers, briefings, synthesis), 'dataset_building' (entity lists, catalogs, extract-all without a URL), 'enrichment' (add fields to records the user already has). Classify the kind first. Only settable when this run creates a new agent (via `agent_name`); against an existing agent it must match the agent's stored use_case.
nimblemcp_nimble_crawl_runStart a web crawl to extract content from multiple pages on a website.Write
Crawl Run
Start a web crawl to extract content from multiple pages on a website. The crawl discovers and visits pages starting from a given URL, following links up to the specified limit. Results are retrieved via `nimblemcp_nimble_crawl_status`. When to use: - You need to extract content from multiple pages on a site automatically. - You want link-following or sitemap-based URL discovery. When NOT to use: - You need content from a single known URL — use `nimblemcp_nimble_extract` instead. - You need results immediately — this is async; first pages complete ~1-2 min after dispatch.
Inputs
urlstringrequired
The starting URL for the crawl. The crawler will discover and visit linked pages from this entry point.
allow_external_linksboolean
Whether to follow links to external domains. Default is False (stay on same domain).default false
allow_subdomainsboolean
Whether to follow links to subdomains of the starting URL. Default is False (stay on same subdomain).default false
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
exclude_pathsstring
Optional list of regex patterns. URLs matching any pattern will be skipped. Example: ["/admin/.*", "/private/.*"]
include_pathsstring
Optional list of regex patterns. Only URLs matching at least one pattern will be crawled. Example: ["/blog/.*", "/docs/.*"]
limitinteger
Maximum number of pages to crawl (default: 100). Higher limits take longer but provide more comprehensive coverage.default 100
namestring
Optional human-readable name for this crawl job.
sitemapstring
How to use the site's sitemap for URL discovery: - "include": Use sitemap in addition to link following (default) - "skip": Ignore sitemap, only follow links - "only": Only crawl URLs found in sitemapdefault include
tool_descriptionstring
Optional description of the tool's purpose.
tool_reasoningstring
Optional reasoning about why this tool was selected.
nimblemcp_nimble_extract_templates_generateKick off generation of a new custom extract template.Write
Extract Templates Generate
Kick off generation of a new custom extract template. Returns a `generation_id`; poll it with `nimblemcp_nimble_extract_templates_generation_status` to get the generated template. On completion the result carries the new `template_name` and `version_id`; use `template_name` with `nimblemcp_nimble_extract_templates_run` to execute it. When to use: - You need a custom extraction template for a site that isn't covered by the pre-built catalog. - You want to generate a template from a natural-language prompt and an example URL. When NOT to use: - A pre-built template already covers the site — use `nimblemcp_nimble_extract_templates_run` instead.
Inputs
promptstringrequired
Natural-language instruction describing what the template should collect (e.g. 'Collect product name, price, and rating from Amazon product pages').
urlstringrequired
Example target URL that helps the generator understand page structure.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
input_schemastring
Optional JSON Schema object describing runtime input parameters.
namestring
Optional desired name for the generated template.
output_schemastring
Optional JSON Schema object describing desired output fields.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_extract_templates_runExecute an extract template against a target URL or set of parameters.Write
Extract Templates Run
Execute an extract template against a target URL or set of parameters. Returns structured data collected from the target page. The params dict must match the template's input_schema — use `nimblemcp_nimble_extract_templates_get` to inspect required fields. Possible issues if no results are returned: - The target URL may be unreachable or behind a login wall. - The page structure may have changed since the template was created. - Required params may be missing or malformed — check the template's input_schema. When to use: - You know the template name and want its structured output inline. When NOT to use: - You want an async run — use `nimblemcp_nimble_extract_templates_run_async` instead.
Inputs
paramsobjectrequired
Runtime input parameters built from the template's input_schema. Include all required keys. Example: {"url": "https://www.amazon.com/dp/B09V3KXJPB"}
templatestringrequired
Template name from nimble_extract_templates_list, nimble_extract_templates_get, or a completed generation.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_extract_templates_update_from_templateCreate a refinement generation that starts from an existing extract template.Write
Extract Templates Update From Template
Create a refinement generation that starts from an existing extract template. Returns a `generation_id`; poll with `nimblemcp_nimble_extract_templates_generation_status` to get the refined template. On success, use the returned `template_name` with `nimblemcp_nimble_extract_templates_run`. When to use: - You want to modify an existing extract template's behavior using natural-language instructions. - You want to extend a template's output schema without rebuilding from scratch. When NOT to use: - You want to build a brand-new template — use `nimblemcp_nimble_extract_templates_generate`.
Inputs
from_templatestringrequired
Template name to refine. A new generation is created from this template.
promptstringrequired
Natural-language refinement instruction (e.g. 'add a ratings field to the output').
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_agents_deleteDelete an agent permanently.Destructive
Agents Delete
Delete an agent permanently. When to use: - You no longer need an agent and want to remove it from the account. When NOT to use: - You just want to stop a specific run — this deletes the agent configuration itself, not a run.
Inputs
agent_idstringrequired
Agent id to delete.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_agents_updateUpdate an existing agent via JSON Patch operations.Destructive
Agents Update
Update an existing agent via JSON Patch operations. When to use: - You need to change specific fields on an existing agent (e.g. its description, goals, or sources) without regenerating it. When NOT to use: - You want to refine an extract template's behavior with natural language — use `nimblemcp_nimble_extract_templates_update_from_template` instead.
Inputs
agent_idstringrequired
Agent id to update.
operationsarrayrequired
JSON Patch operations to apply. Each item: {op: 'add'|'remove'|'replace'|'move'|'copy'|'test', path: "/field", value: ...}. Example: [{"op": "replace", "path": "/description", "value": "New"}]
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_crawl_terminateCancel a running or queued crawl job.Destructive
Crawl Terminate
Cancel a running or queued crawl job. When to use: - You want to stop a crawl that is no longer needed before it completes. When NOT to use: - The crawl has already succeeded or failed — it cannot be cancelled in a terminal state.
Inputs
crawl_idstringrequired
The unique crawl identifier returned by nimble_crawl_run.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.
nimblemcp_nimble_extract_asyncStart an asynchronous URL extraction.Destructive
Extract Async
Start an asynchronous URL extraction. Returns immediately with a task ID. Poll `nimblemcp_nimble_task_results` to retrieve the extracted content when ready. When to use: - You want to extract a URL without blocking while it renders. - The page is complex or slow to load. When NOT to use: - You need the result inline — use `nimblemcp_nimble_extract` for synchronous extraction.
Inputs
urlstringrequired
The URL to extract content from.
callback_urlstring
Optional webhook URL to receive results when extraction completes.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
countrystring
Country code for geo-targeted results (e.g., "US", "GB", "IL"). Defaults to "US".
driverstring
Optional browser driver tier for web extraction: - "vx6": fast HTTP requests, no JavaScript â static pages and APIs - "vx8": headless browser with JavaScript rendering â dynamic pages and SPAs - "vx8-pro": headful browser with JavaScript rendering â complex interactions - "vx10": stealth headless browser â sites with strict bot mitigation - "vx10-pro": stealth headful browser â heavily bot-protected sites If omitted, the backend selects. Escalate explicitly for JavaScript-heavy pages ("vx8"/"vx8-pro"), or for sites with strict bot mitigation ("vx10"/"vx10-pro"), when a default extraction returns empty or truncated content.
localestring
Language/locale code (e.g., "en", "fr", "de"). Defaults to "en".
output_formatstring
Content output format: - "markdown": Markdown-formatted content (default) - "plain_text": Plain text without formatting - "simplified_html": Clean HTML without scripts/stylesdefault markdown
tool_descriptionstring
No description.
tool_reasoningstring
No description.
waitstring
Optional wait time in milliseconds for JavaScript-rendered pages.
nimblemcp_nimble_extract_templates_run_asyncStart an asynchronous extract template run.Destructive
Extract Template Run Async
Start an asynchronous extract template run. Returns immediately with a task ID. Use this for long-running template extractions or batch processing. Retrieve results later with `nimblemcp_nimble_task_results` using the returned task ID. When to use: - You want to run an extract template without blocking while it completes. - You have a long-running or batch extraction where you will poll for results. When NOT to use: - You want the result inline — use `nimblemcp_nimble_extract_templates_run` for a synchronous run.
Inputs
paramsobjectrequired
Runtime input parameters matching the template's input_schema. Example: {"url": "https://www.amazon.com/dp/B09V3KXJPB"}
templatestringrequired
Extract Template name to execute. Get available templates from nimble_extract_templates_list.
callback_urlstring
Optional webhook URL to receive results when the run completes.
contextstring
Explain why you are calling this tool and how it fits into the user's overall goal. This parameter is used for analytics and user intent tracking. YOU MUST provide 15-25 words (count carefully). NEVER use first person ('I', 'we', 'you') - maintain third-person perspective. NEVER include sensitive information such as credentials, passwords, or personal data. Example (20 words): "Searching across the organization's repositories to find all open issues related to performance complaints and latency issues for team prioritization."
conversation_idstring
Echo the conversation_id from the server's previous response. The server provides it on the first call â never invent one, and do not issue parallel tool calls until you have it.
tool_descriptionstring
No description.
tool_reasoningstring
No description.

Setup

Install the SDK

Set your credentials

Create the Nimble MCP connection

Create an API key in Nimble

Add the API key to Scalekit

Authorize a user and make your first call

Tools

Call it

Returns

Agent Templates Get

Inputs

Agent Templates List

Inputs

Agents Get

Inputs

Agents List

Inputs

Agents Run Result

Inputs

Agents Run Status

Inputs

Agents Runs List

Inputs

Crawl List

Inputs

Crawl Status

Inputs

Extract

Inputs

Extract Templates Generation Status

Inputs

Extract Templates Get

Inputs

Extract Templates List

Inputs

Map

Inputs

Search

Inputs

Task Results

Inputs

Agents Create

Inputs

Agents Run

Inputs

Crawl Run

Inputs

Extract Templates Generate

Inputs

Extract Templates Run

Inputs

Extract Templates Update From Template

Inputs

Agents Delete

Inputs

Agents Update

Inputs

Crawl Terminate

Inputs

Extract Async

Inputs

Extract Template Run Async

Inputs

Related: other AI connectors