Keep and reuse results
No company's docs show a full fit yet. The listings below do the job with a limit, or not at all.
What counts as done
The provider's terms permit storing returned text and URLs for the intended use, found in the terms within a minute of reading, and the API exposes nothing that silently caches on the agent's behalf.
How agents ask for it: save these results to the database
, build a dataset from searches
, cache results so we do not pay twice
Does the job, with a limit 5
It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.
| Listing | Why |
|---|---|
| Brave Search API + MCP B Brave Software, Inc. | Terms forbid storing, caching or building a database of Search Results beyond transient operation, yet one user found Brave the only provider whose plan (Enterprise storage rights) lets them store results. source |
| Cloudflare Web Search API B Cloudflare, Inc. | Cloudflare's pages say nothing about storage; Ceramic's terms restrict caching beyond real-time display per a first-hand account, and Exa's or Linkup's terms apply when routed to them. source |
| Mojeek Web Search API F Mojeek Limited | Startup may cache results only for an hour to serve the same user; Business and Enterprise plans carry storage rights permitting indefinite retention, and results may not feed a search index. source |
| Perplexity Search API BB Perplexity AI, Inc. | The overview says results are data for your own code to filter, store or hand to a model, but the API terms were not read. source |
| Tavily API + MCP A AlphaAI Technologies Inc. dba Tavily | The terms say nothing against storing Outputs, so nothing forbids it, but nothing grants it either, and Tavily may use Customer Input and Outputs to train models. source |
Couldn't check 17
The docs were closed to us, or silent. These carry no mark either way.
| Listing | Why |
|---|---|
| Arcmira B Arcmira, Inc. | Terms of service not opened in this pass; the docs read say nothing about storing passages. |
| DeepSeek Search and Fetch API C Hangzhou DeepSeek Artificial Intelligence Co., Ltd. | Terms of use not read. |
| Exa API + MCP B Exa Labs Inc. | exa.ai/terms needs JavaScript, so the storage clause was not read. |
| Grep AI B Parcha Labs, Inc. | Terms not opened; the report belongs to the run and the OpenAPI document says nothing about reuse. |
| Liner Search MCP E Liner | Terms not read in this pass. |
| Linkup B Linkup Technologies SAS | www.linkup.so/terms returned 404 and the docs say nothing about storing results. |
| NewsCatcher B NewsCatcher, Inc. | Terms not read. |
| Parallel Search and Task APIs BB Parallel Web Systems Inc. | Customer Terms not read in this pass. |
| SearchAPI.io D SearchApi, LLC | Terms not read. |
| SerpApi B SerpApi, LLC | Terms not read. |
| Serper F Serper | Terms not read. |
| SERPHouse E Patoliya Infotech LLP | Terms not read in this pass. |
| SerpKite BB Anastasia Klimova, trading as SerpKite | Terms not read in this pass. |
| SpecStream B SpecStream SAS | Terms not read. |
| TinyFish Search B Tiny Fish, Inc. | Terms not read in this pass. |
| Valyu BB Valyu.Network Ltd | Terms not read in this pass. |
| You.com APIs BB SuSea, Inc., d/b/a you.com | Terms not read. |
Doesn't do this job 4
Listings in the category that do something else, with the reason.
| Listing | Why |
|---|---|
| Web Search Tool on Amazon Bedrock AgentCore BB Amazon Web Services, Inc. (regional AWS entities by account location) | Acceptable use: you may not extract, store or reproduce content from Search Results in bulk, and must retain and display citations in any output. source |
| Grounding with Bing Search B Microsoft Corporation | Terms: Customer may copy, store or cache Bing Services Output only as an integrated part of its own work product and may not otherwise store or cache it, or train AI models on it. source |
| Fetch (MCP reference server) B Model Context Protocol, a Series of LF Projects, LLC | Not a search provider; what may be kept depends on each fetched site's terms, which the server does not read. |
| Kagi Search API C Kagi Inc. | API terms: do not store, cache or build a database of Results beyond transient caching, and delete cached Results on termination. source |
What stops agents finishing
- terms restrict caching to real-time display (Ceramic)
- terms require deletion on termination and allow only transient caching (Kagi)
- acceptable-use terms bar bulk storage (AWS AgentCore)
Around it
Before: search.ground-fact. After: memory.store, data-tables.append-row.
Where this comes from
Weight 4 of 100 in Web search APIs: 3 of 44 accounts: a well-known agent builder asking about resyndication, a user who could only find Brave permitting storage, and Kagi, AWS and Ceramic terms read for this map
The same as data: /api/v1/use-cases/search.store-results.json, or the MCP tool get_use_case with id search.store-results. Other jobs in this category: /api/v1/niches?category=search.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.