Linkup
Web search API for AI agents that returns cited, full-text sourced answers in one call.
If your agent's worst failure is a confidently wrong answer, Linkup belongs on the shortlist — it publishes reproducibility benchmarks (92% Verified SimpleQA F-score, 61% SealQA-0) and offers BYOC deployment that most rivals won't commit to in writing. The three-endpoint model costs you real integration thinking up front, and the free allotment ends at 4,000 queries. Buy it for accuracy and data residency; look at Exa when you actually need semantic discovery.
Verified 2h ago · liveness 79/100 · cite: rightaichoice.com/tools/linkup
- Enterprise AI apps that need SOC 2 Type II, ZDR and regional data residency evidence before production
- Legal and regulated-industry AI platforms that must ground answers in citable web evidence
- Agent frameworks and computer-use agents needing cited web access inside latency budgets
- Research copilots running multi-step deep research where 1-10 minute async jobs are acceptable
- Hobby projects on a tight budget — the free allotment stops at 4,000 queries
- Semantic similarity or "find me things like this" discovery workloads, where Exa fits better
- Latency-critical chat turns that cannot absorb a 1-10 minute asynchronous Research call
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Linkup if you need semantic similarity discovery rather than cited factual retrieval, or if your agent loop cannot tolerate anything slower than a sub-200ms evidence fetch and you're not using Linkup Flash.
Research calls run $0.25 to $2.50 per request, so a deep-research agent making hundreds of calls per day burns through the free 4,000 queries far faster than Fetch-only usage would.
Pay-as-you-go makes Linkup cheap to start and unpredictable to scale — a Fetch-only prototype costs fractions of a cent per call, but a Research-heavy agent at $0.25-$2.50 per request can cost more per day than a flat-rate search vendor. Enterprise adds custom pricing, private indexes, ZDR, and SLA coverage. Against Tavily's subscription tiers Linkup is more expensive at high Research volume; against Exa it competes on accuracy and compliance rather than price.
In short
Linkup — Web search API for AI agents that returns cited, full-text sourced answers in one call. Best for Enterprise AI apps that need SOC 2 Type II, ZDR and regional data residency evidence before production, Legal and regulated-industry AI platforms that must ground answers in citable web evidence, Agent frameworks and computer-use agents needing cited web access inside latency budgets. Paid, in a currency we have not confirmed — see the pricing table for the vendor’s own figures.
No head-to-head comparison for Linkup yet — the alternatives below are the closest matches we list.
What's new in Linkup
Checked 2 days agoAcross the latest 10 updates: 2 launches, 3 community discussions and 5 news mentions.
Aleph Alpha x Linkup case study: sovereign AI search in the EU
Aleph Alpha AI Solutions engineering lead cites Linkup for GDPR compliance, zero data retention and no EU-external processing at sufficient index quality.
Linkup post: sovereign AI is a control problem
Linkup argues sovereignty is about controlling the rules your AI runs under, and frames the questions buyers should ask vendors.
Migration guide: Exa to Linkup in under an hour
Linkup publishes a step-by-step migration from Exa, pitched on accuracy, lower cost and enterprise compliance.
Linkup releases SPARSEUP open sparse embedding model
SPARSEUP, Linkup's open-source sparse embedding model, is live with weights on Hugging Face (linkup-sparseup-embed-v1).
Linkup Flash (Alpha) search API ships sub-200ms
Linkup Flash (Alpha) is a web search API at under 200ms median latency, measured client-side, aimed at grounding agents.
Linkup GPU pipeline tuning: 37k to 200k tokens/sec
Linkup engineering reports a same-day gain from 37,000 to over 200,000 tokens per second on identical GPU and model.
Linkup publishes enterprise web search API buyer's guide
Linkup's 2026 buyer's guide argues compliance review, benchmark accuracy and latency at scale matter more than per-query price.
Migration guide: Tavily to Linkup in under an hour
Linkup documents migrating from Tavily, citing higher accuracy, lower cost and enterprise compliance as drivers.
Linkup on data enrichment for AI agents
Linkup advocates agent-time field fetching from the live web with a source citation per attribute.
Linkup: knowledge reconciliation still needs a grounding API
Linkup argues self-updating agents still require a real-time web search API as their evidence source.
What people actually say about Linkup — is it worth it?
We scanned public community sources for Linkup on Aug 30, 2026 and could not establish that the discussion we found is about this tool rather than something else sharing its name. Our own analysis of that scan says the posts were off-subject. Rather than publish a sentiment score built on the wrong subject, we publish nothing here and re-run the scan.
Viability Score
How well maintained and how widely used is Linkup? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Search endpoint with Fast, Standard and Deep modes returning synchronous results in 1-3 seconds
- Fetch endpoint for synchronous web page content extraction in under 2 seconds
- Research endpoint running asynchronous deep research jobs in 1-10 minutes
- Linkup Flash (Alpha) delivering web evidence at sub-200ms median client-side latency
- Sourced, cited answers with full-text snippets from trusted web sources
- Mercator crawler turns the open web into AI-navigable, indexed data
- SPARSEUP open sparse embedding model published on Hugging Face (linkup-sparseup-embed-v1)
- Configurable source selection, index tuning and freshness controls
- Private Index deployment over access-controlled proprietary data
- Bring Your Own Cloud (BYOC) deployment running the indexing layer in your infrastructure
- Zero data retention and SOC 2 Type II with SLA-backed 99.9% uptime on Enterprise
- 92% F-score on Verified SimpleQA benchmark for search-augmented factual accuracy
- 61% SealQA-0 score on deep research benchmarks
- Time machine: search the web as it was at a chosen date for RL backtesting
- OpenAI SDK wrapper, LangChain and CrewAI integration in one line of code
About Linkup
Linkup is a web search API for AI agents, RAG pipelines and research copilots that need sourced, cited answers rather than unsupported model guesses. Instead of a single search box it splits the job across three endpoints: Fetch retrieves content from a URL synchronously, Search returns results across Fast, Standard and Deep modes, and Research runs asynchronous deep research jobs. A newer Flash tier pushes web evidence to a sub-200ms median client-side latency, aimed at grounding agents that can't wait seconds for a tool call. The pitch is production reliability rather than developer-experience gloss. Linkup runs its own Mercator crawler, publishes a 92% F-score on Verified SimpleQA plus a 61% SealQA-0 deep research score, and keeps SOC 2 Type II on every plan. Deployment is where it separates from thin API wrappers: a Private Index built over access-controlled proprietary data, and Bring Your Own Cloud so the whole indexing layer runs inside your infrastructure. The 2026 Aleph Alpha case study is the clearest proof point, citing GDPR compliance and no EU-external processing as the reasons a sovereign AI vendor standardized on it. Integration is deliberately shallow — a one-line drop-in via the OpenAI SDK wrapper, plus LangChain and CrewAI support — and the vendor publishes step-by-step migration guides from both Exa and Tavily, claimed to take under an hour. Pricing is pay-as-you-go with the first 4,000 queries free, then per-request rates across Fetch, Search (Fast/Standard/Deep) and Research. Enterprise adds custom indexes, a dedicated refresh rate, ZDR and SLA terms. The vendor positions Linkup against Perplexity's Sonar API on raw cited sources plus zero data retention, against Exa and Tavily on accuracy and compliance, and against in-house crawling on the fact that you stop maintaining a pipeline. For teams where the failure mode is confident fabrication rather than slowness, that trade is the whole argument.
Behind the Verdict
Pick Linkup when web evidence has to be defensible. Legal, financial and EU-regulated AI products are the clearest fits — the Aleph Alpha deployment is a live example of a vendor choosing Linkup specifically because nothing left EU jurisdiction and zero data retention held up to review. If your compliance team asks who touches the data and where, that answer is easier to give here than with most search APIs. We'd also reach for it in agent loops. Search returns synchronously across three depth modes, and Flash cuts the evidence path to a sub-200ms median client-side latency — that's the difference between a tool call an agent can afford on every turn and one it has to schedule. For multi-step research where a 1-10 minute async job is acceptable, the Research endpoint earns its keep at $0.25-$2.50 per request. Where it bites: choosing between Fetch, Search and Research is a design decision you make once and live with. Teams coming from a single universal search endpoint often under-scope that. And the free tier stops at 4,000 queries, so a weekend prototype can hit the paywall fast. Compared to Exa, the split is workload, not quality. Exa is built for semantic similarity and discovery; Linkup is built for retrieving grounded facts with citations. If you're doing "find me companies like this one," the wrong tool will feel wrong immediately. Against Tavily, Linkup leans on accuracy and enterprise compliance rather than tutorial breadth — and it now publishes migration guides for both, which tells you where it thinks the switchers are coming from. Practical caveat on the benchmarks: 92% Verified SimpleQA and 61% SealQA-0 are strong, but they are Linkup's own published methodology. Run your own eval set against your domain before signing anything, and put your two
Researching Linkup? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Linkup actually fits — and what changes day-one when you adopt it.
You wire the Search endpoint into your LangChain retrieval step for cited case-law lookups in 1-3 seconds, and route long multi-hop questions to the Research endpoint as background async jobs that write back to the answer cache.
Outcome: Your copilot returns sourced, cited legal analysis without hallucinated statutes, and you can point enterprise reviewers at the 92% Verified SimpleQA F-score and SOC 2 Type II compliance.
You call Search at account-signup time to enrich the CRM record with company data and buying signals, then hand live prospect intel to the drafting agent.
Outcome: Reps see fresh company context before their first call, and your enrichment pipeline stays within the per-request pricing because Search calls run $0.005-$0.006 each.
You deploy Linkup's indexing layer inside your own cloud via BYOC, build a private index over access-controlled internal data, and enable Zero Data Retention under an Enterprise agreement.
Outcome: Your AI agents retrieve from internal and web sources with nothing leaving the environment, backed by a 99.9% uptime SLA and IP whitelisting.
Use Cases
- Build a real-time AI copilot that answers complex technical questions with up-to-date web data.
- Enrich CRM and lead generation workflows with corporate intelligence from LinkedIn and premium sources.
- Automate invoice recovery by researching debtor information and drafting tailored messages.
- Create a research assistant that retrieves in-depth, cited analysis for legal or financial domains.
- Power a voice agent that searches the web for live information on products, services, or events.
- Run agentic corporate reputation monitoring with live web context.
- Provide secure enterprise AI search for regulated industries like railway operations.
- Train reinforcement learning search agents on the same retrieval layer you ship to production.
Models Under the Hood
as of 2026-10-16
Limitations
- The Research endpoint runs asynchronously with 1-10 minute latency, unsuitable for real-time use, while Fetch (<2s) and Search (1-3s) are synchronous but still add seconds; Linkup Flash is the sub-200ms option for agent loops that cannot wait.
- Free usage is capped at 4,000 queries, and Zero Data Retention, custom/private indexing, premium features and higher index refresh rates require the Enterprise plan.
- Per-request costs are tiered ($0.001-$0.006 Fetch, $0.005-$0.006 Search, $0.25-$2.50 Research), so spend scales with Search and Research call volume.
- The evidence shows only an API (REST via Python, Node.js, cURL) with no prebuilt app UI, so it is aimed at developers wiring search into AI systems.
as of 2026-10-01
Verification history
We have re-verified Linkup 22 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-checked, vendor evidence unchanged
- — re-checked, vendor evidence unchanged
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 22 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Linkup tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Pay as you go
First 4,000 queries free, then usage-based
Ideal for
Teams prototyping or running moderate production volume who want to test on 4,000 free queries before committing to enterprise terms.
What this tier adds
Starting tier: 4,000 free queries, then per-request pricing across Fetch ($0.001-$0.006), Search ($0.005-$0.006), and Research ($0.25-$2.50). SOC 2 Type II included.
Enterprise
Custom
Ideal for
Regulated enterprises and legal AI platforms that need ZDR, private indexes, residency by region, and an SLA before they can ship.
What this tier adds
Adds custom pricing models, personalized private indexes, dedicated index refresh rate, private environment deployment, ZDR, IP whitelisting, SLA coverage, and dedicated support.
Where the pricing makes sense
The company stage and team size where Linkup's pricing actually pencils out — and where peers do it cheaper.
Pay-as-you-go makes Linkup cheap to start and unpredictable to scale — a Fetch-only prototype costs fractions of a cent per call, but a Research-heavy agent at $0.25-$2.50 per request can cost more per day than a flat-rate search vendor. Enterprise adds custom pricing, private indexes, ZDR, and SLA coverage. Against Tavily's subscription tiers Linkup is more expensive at high Research volume; against Exa it competes on accuracy and compliance rather than price.
Setup time & first value
How long it actually takes to get something useful out of Linkup — broken out by persona, not the marketing-page minute.
Fetch and Search endpoints: an afternoon. The OpenAI SDK wrapper is a one-line drop-in, so a developer with an API key can run a first cited query in under an hour, and the vendor's own Tavily and Exa migration guides claim the switch takes under an hour. Research endpoint workflows: a few days, because you need job queuing for the 1-10 minute async calls. Enterprise BYOC / private index: weeks,
Switching to or from Linkup
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From Tavily: Linkup publishes a Tavily-to-Linkup migration guide claiming the switch takes under an hour.
- →From Exa: Linkup publishes an Exa-to-Linkup migration guide claiming the switch takes under an hour.
- →From Perplexity Sonar API: Linkup positions itself as a Sonar alternative for RAG, citing raw cited sources plus zero data retention.
- →From a self-built crawler: replace your scraping pipeline with the Mercator-backed Search and Fetch endpoints.
- ↗To Exa: move if your workload is semantic similarity or discovery rather than cited factual retrieval.
- ↗To Tavily: move if you want a large library of tutorials and a simpler single-endpoint mental model over three separate endpoints.
- ↗To Perplexity Sonar API: move if you want a synthesized answer layer rather than raw cited sources.
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “Linkup”, and we withheld 6: 6 could not be judged, because “Linkup” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Linkup.
Official links
Tools that pair well with Linkup
Common stack mates teams adopt alongside Linkup, with the specific reason each pairing earns its keep.
You.com
You.com is a web search API for AI agents — real-time results, clean page extraction, and cited answers, priced per call.
Mixedbread AI
Hosted retrieval API that returns cited evidence — file, page and passage — to AI agents instead of raw search hits
Reprompt
Deep research AI agents that verify the live status and real ownership of physical businesses, returning cited answers instead of rows from a stale POI file.
Alternatives to Linkup
View all
You.com
You.com is a web search API for AI agents — real-time results, clean page extraction, and cited answers, priced per call.
Mixedbread AI
Hosted retrieval API that returns cited evidence — file, page and passage — to AI agents instead of raw search hits
Frequently Asked Questions
Categories
Best-of guides
Used Linkup? Help shape our editorial sentiment research.