ae_tavily_research
根据问题生成带来源的研究报告,支持模型、引用格式、报告长度、域名偏好、结构化输出和附件。本工具异步执行:先返回AE任务ID,再通过ae_get_result查询;查询同一任务不会重新提交研究。预占按所选模型官方上限,最终按实际用量结算。
AgentEarth Tool Catalog
Connect once to Google, Bing, Tavily, Exa, Firecrawl and more for web search, full-site crawling and structured extraction at lower cost.
Tool List
ae_tavily_research
根据问题生成带来源的研究报告,支持模型、引用格式、报告长度、域名偏好、结构化输出和附件。本工具异步执行:先返回AE任务ID,再通过ae_get_result查询;查询同一任务不会重新提交研究。预占按所选模型官方上限,最终按实际用量结算。
ae_law_get_item_content
按一部法的名称和一条的条号,取这一条的原文、发文字号、时效和法宝链接。不是法规列表,也不是这一条历年怎么改。
ae_law_search_article
【工具功能】根据用户输入的案情、法律问题或一段自然语言,对北大法宝法规库按法条正文级别进行语义检索,返回语义相关的法条原文,及法规标题、法条编号、发文字号、时效性、效力位阶、制定机关、公布日期、施行日期、北大法宝链接等结构化信息,并支持按法规库、时效性、制定机关、施行日期范围进行筛选过滤。 【何时使用】当用户提出与法条内容相关的任何问题时应优先使用,包括但不限于:查询某一法律问题违反的法条依据、核实某部法规某条的具体内容、检索特定日期之后发布的法条正文、需要根据权威法条内容支撑法律分析、合规判断和文书撰写等。 【何时不要使用】不要使用本工具查询法规的基本信息,此时应该优先使用"search_laws"工具。不要使用本工具查询法律法规之外的内容,如司法案例、裁判文书、法学期刊、学术论文、新闻资讯、企业内部制度或地方政策文件解读等非法规库收录的内容。 【结果可信级别】结果可信级别为高,数据来源于北大法宝权威法规数据库。返回结果是空列表表示「这组查询条件下没有命中」,是有效查询结果。每条结果请一并给出citation,便于用户核对。该工具为语义检索,返回结果与查询意图可能存在偏差,涉及正式法律意见或对外引用时,请人工复核法规内容。
ae_brightdata_unlock
通过Bright Data Web Unlocker请求指定网页或公开数据接口,可指定出口国家、HTTP方法和请求体。返回HTML或Markdown;需自行提供目标URL与参数,不能自动发现业务接口。对会修改目标系统数据的请求,调用者须事先取得授权。
ae_api_sports_football_get_fixtures
Search for football matches by date, league, team, or live status using API-Sports.
ae_bid_get_price_trends
Query historical winning bid price trends, including price statistics (maximum, minimum, average, median) and price timelines. Suitable for price analysis, budgeting, and cost control. Requires at least one filter among brand, model, or product. Dates must use YYYY-MM-DD format.
ae_firecrawl_agent_status
Check the progress or retrieve the completed result of an existing Firecrawl agent job. Input: `id` (required string), the job ID returned by `firecrawl_agent`. It is the only accepted parameter; the schema does not impose a UUID format. Returns: a JSON object that may contain `id`, `success`, `status`, `progress`, `data` when completed, and `error` when unsuccessful. `success` describes the API call, not whether the research has finished. `processing` means the job is still running; `completed` and `failed` are terminal states. Failed jobs may return an incomplete best-effort `partial` result and `partialSchemaValid`; partial output must not be presented as a completed answer. Optional metadata includes `sources`, `message`, `mode`, `model`, `threadId`, `threadTurn`, `conversationId`, `suggestions`, `creditsUsed`, `expiresAt`, `stopReason`, `notice`, `exchange`, `pendingApproval`, and `agent_hints`. Availability varies, and some values may be null. `credit_limit_reached` indicates that a failed run reached its credit limit. Results are no longer readable after their expiration. Asynchronous behavior: the underlying research runs asynchronously, but each invocation of this tool performs one status/result lookup. It does not submit a job, wait until completion, or automatically poll. While the status is `processing`, check again after 15–30 seconds if time permits. Typical runs take roughly 1–3 minutes; complex research can take 5 minutes or more. These durations are estimates, not guarantees, and elapsed time alone does not establish failure. Stop completion polling on `completed` or `failed`. Limitations: read-only lookup for an existing job. It cannot start, modify, cancel, or resume research, accept provider terms, or approve paid calls. Approval-related fields only report information; no approval input is supported here. No prompt, polling-interval, or timeout parameter is exposed.
ae_linkup_research
Start a deep, autonomous web research task with Linkup to get synthesized, cited answers for complex questions.
ae_bid_get_company_profile
Query enterprise registration details, business profile, industry classification, organization scale, and bidding statistics using a company name, ID, or detail URL.
ae_dataforseo_ai_optimization_keyword_data_search_volume
Get estimated search volume data for keywords as used in AI large language models (LLMs) via DataForSEO.
ae_dataforseo_labs_google_relevant_pages
Analyze a domain's page-level SEO performance using DataForSEO. Retrieve ranking distributions and estimated monthly traffic from organic and paid search for specific web pages.
ae_you_post_research
Generate a detailed, cited answer to complex questions using multi-step research.
ae_bid_search_proposed_projects
Search proposed projects (in initiation or approval stages) with filtering by keywords, tendering entity, geographical regions, date range, approval status, and budget amount.
ae_brightdata_run_scraper
Run a Bright Data scraper to collect records for URLs or queries. Provide dataset_id from search_scrapers, a method copied exactly from the scraper's collection_methods, and a nonempty input array with one object per URL or query. Check get_scraper_details for that dataset_id and method first, and build each input object from sample_input or input_fields. collect_by_url returns one record per input; discover_by_* methods return up to limit_per_input records per input. The optional limit_per_input applies only to discovery methods, accepts integers from 1 to 1000, and defaults to 10. Every returned record is billed. The call waits up to 45 seconds and returns collected records if complete, or snapshot_id with status "running" if processing continues. Use get_scraper_progress to check the run, then get_scraper_results when ready. Records with an error field explain failed inputs; other record fields depend on the selected scraper and method.
ae_dataforseo_labs_google_serp_competitors
Discover competitors ranking for specific keywords using DataForSEO. Get SERP positions, visibility scores, and estimated traffic volumes for domains competing on your target terms.
ae_dataforseo_on_page_lighthouse
Run a Google Lighthouse audit via DataForSEO to measure web page quality, performance, and accessibility.