The Router — the part that makes it effortless
State a goal. The router works out what intelligence is required, where to retrieve it, and how to obtain it — across the state-of-the-art AI labs, licensed data providers and open-web sources on the market. You get their collective coverage without understanding any of them semantically, commercially, or technically.
18 AI labs · 23 licensed data providers · 19 signal sources — and anything else with an API
What the router decides
These are the judgment calls that normally cost a RevOps hire, a data analyst and an engineer. In Hyves they are just how the machine works — made per task, in milliseconds, with fallbacks.
The router decomposes a goal into its actual needs. A market map needs registries and review sites; a verified email needs a waterfall; an opener needs the dossier and a model that can write in your voice. Nobody sits down to spec this — it falls out of the mission.
Per need, it picks the best source on the market — not the one you happen to have a contract with. Registry data from registry sources, technographics from technographic specialists, reasoning from the lab that is strongest at that step. Strengths, coverage, freshness and cost all weigh in.
Then it speaks whatever the source speaks — REST API, MCP server, native connector. Auth, rate limits, retries, pagination, schema quirks and fallbacks are absorbed here, and what comes back is verified before it is allowed to touch your memory.
Why this is the moat
The market's best providers are only useful if someone on your team understands each one — what its fields mean, whether it is worth the money, and how to keep its integration alive. That someone is usually three someones.
The router carries all three kinds of understanding, for every provider at once. Your team states goals; the whole market answers.
You never learn each provider's private language — what an “intent topic”, a “technographic install” or a “verified contact” means inside that one source's schema. The router speaks all of them and normalizes every answer into one memory, so your team only ever reads one vocabulary.
You never rank vendors, decode credit pricing, or know which enrichment source is worth it for which country and segment. The router carries that judgment: it routes each request to the source that earns it, cross-checks the answer, and quietly stops using what underperforms.
No API keys to juggle, docs to read, SDK versions to pin, webhooks to babysit. If a system exposes an API or an MCP server, the router can drive it — and when a provider degrades mid-task, it falls back to the next one instead of falling over.
The routable market
Browse the intelligence market the router aggregates across — frontier models, licensed datasets, and the open-web sources where buying signals are born. Some of it runs natively in Hyves today; the rest is reachable the moment a mission needs it, through API or MCP. You never need to know which is which — that is the point.
GPT reasoning and Whisper speech
Claude for judgment-heavy steps
Gemini reasoning, Veo video
Llama for scaled workloads
European open-weight models
Frontier open-weights at swarm scale
GLM frontier models
Cost-efficient deep reasoning
Grok with live X signal
Kimi for very long documents
Enterprise embeddings and rerank
Search-grounded answers with citations
FLUX image generation
Diffusion image pipelines
Generative video for campaigns
Natural voice for calls and audio
Serverless open-model inference rail
LPU inference when latency matters
Person and company records at scale
Firmographic and technographic enrichment
Funding rounds and company profiles
European startup and investor data
Company registries, worldwide
Licensed B2B contact database
Contact coverage with intent hints
GDPR-first European contact data
Direct dials and verified emails
Hard-to-find contact coverage
Domain patterns and email finding
GDPR-clean email enrichment
Waterfall email enrichment
Multi-source contact waterfall
Verified B2B email discovery
Email discovery from profiles
Deliverability checks before send
List hygiene at send time
Full tech-stack history
Live technology detection
Traffic and digital footprint
Search visibility signals
E-commerce store intelligence
Web-scale collection: registries, reviews, careers
Agentic web search for live facts
Neural search over the open web
SERP results at API speed
Structured extraction from any page
Software reviews and switch signals
Incumbent pain, written by their users
Customer sentiment at scale
Workforce sentiment and attrition hints
Leadership hires and company posts
Hiring spikes as roles go live
Launches, chatter, mentions
Unfiltered practitioner signal
Launch radar for tech vendors
OSS adoption and engineering signal
Public tenders and RFPs, daily
Dutch registry filings at source
UK filings straight from source
Global news firehose, machine-readable
This catalog is directional, not a ceiling — any intelligence source with an API or an MCP server is routable, and we will tell you straight how your market would route. Looking for the business apps Hyves keeps in sync — CRM, commerce, finance? That is the integration swarm, on the integrations page.
AUTO-FALLBACKS
A degraded provider is routed around, not waited on
VERIFICATION WATERFALL
Sources cross-check each other before anything enters memory
APPROVAL-GATED
Everything outward-facing holds for a human
COST-AWARE
Routed on merit per task, never vendor loyalty
Bring us your market. We will run the router on it and show you the intelligence it aggregates — sources, signals and all.