Skip to content
GridRouterhome

Search

Search providers, capabilities and pages

Sign inGet your free API key

443 vendorsThe OpenRouter for go-to-market data

One key for every GTM data tool.

Enrichment, email, phone, intent, ads, SEO and CRM calls through one API and one MCP server. See the price before every call, and measured hit rates as traffic comes in. Routing is free, and calls on your own vendor keys are never marked up.

Get your free API keyBrowse the catalog

Free while we launch. Bring your own vendor keys.

curl https://api.gridrouter.io/v1/run/people.email.find \
  -H "Authorization: Bearer $GRID_API_KEY" \
  -H "X-Grid-Max-Cost: 0.05" \
  -H "Idempotency-Key: lead-8841" \
  -d '{"input": {"first_name": "Ada", "last_name": "Lovelace",
                 "domain": "example.com"}}'
{
  "object": "call",
  "endpoint_id": "hunter/email.find",
  "provider": "hunter",
  "status": "succeeded",
  "hit": true,
  "cost_micro": 30000,
  "credential": "managed",
  "data": { "email": "ada@example.com", "confidence": 94 },
  "attempts": [{ "provider": "hunter", "outcome": "hit", … }]
}

Routes to 443 data vendors, on one bill

See all 443 providers

Logos are trademarks of their owners and identify the providers GridRouter routes to.

Vendors listed
443
Each with sourced docs, pricing and auth
Capabilities
59
One input and output schema each
Categories
22
From enrichment to ads and SEO
Callable endpoints
0
Scores appear as calls are routed
Catalog checked

Every GTM data job, one schema each

59 capabilities across 22 categories. Each has one input shape and one output shape, whichever provider answers.

Browse the catalog

Search, see the price, then call

  1. 1Step 1 of 3

    Search the catalog

    Ask for the job, not the provider. Every capability has one id and one input shape, whichever provider answers it.

    GET /v1/catalog/endpoints?q=email

hunter/email.find     people.email.find  per_success  $0.03
hunter/email.verify   email.verify       per_call     $0.012
…

  2. 2Step 2 of 3

    See the price first

    A quote returns the worst-case cost of a call before anything is charged. Measured fill rate and accuracy sit beside each endpoint as traffic comes in.

    POST /v1/quote/hunter/email.find
{"input": {"first_name": "Ada", "last_name": "Lovelace",
           "domain": "example.com"}}

{"endpoint_id": "hunter/email.find",
 "quote_micro": 30000, "pricing_model": "per_success"}

  3. 3Step 3 of 3

    Call, run, or send a list

    One call for a live lookup, a routed run with a cost cap, or up to 10,000 rows in the background with a webhook when they're done.

    POST /v1/lists
{"target": {"capability": "people.email.find"},
 "rows": [{"first_name": "Ada", …}, …],
 "webhook_url": "https://example.com/hooks/grid"}

{"object": "list", "id": "list_…", "status": "running",
 "total": 5000, "done": 0, "hits": 0, "cost_micro": 0}

What's in GridRouter

Routing is the start. Every call also gets logs, billing checks, a private cache and alerts.

Product overview

Built for agents that spend money

Give an agent one token instead of twelve API keys, and a budget it can't exceed.

  • MCP server

    Claude Code, Cursor and any MCP client get catalog_search, quote, call, run and list_create as tools, with prices on every endpoint before the agent spends.

  • Agent keys

    Each agent gets its own key and budget. Revoke one agent without rotating anything else.

  • Budgets

    Cap an agent, a key or the whole workspace per day, week, month or in total. Every call lands in the logs with provider, price, latency and outcome.

Create an agent with a $25 monthly budget
POST /v1/agents
{"name": "inbound-router", "budget_micro": 25000000,
 "period": "month"}

{"object": "agent", "id": "agt_…", "name": "inbound-router",
 "budget_micro": 25000000, "period": "month",
 "key": {"prefix": "grid_live_…", …}}

Scores come from traffic, not from vendors

Every number is recomputed from routed calls and shows its sample size. Endpoints without enough calls say so instead of showing a score.

Reliability
Share of calls that returned a well-formed response, with the sample size beside it.
Fill rate
Share of calls that found what was asked for, by each capability's own hit definition.
Accuracy
Share of hits confirmed correct downstream: bounces after send, dials that connected, profiles re-checked.
Speed
p50 and p95 latency from our edge, per provider, over 30 days.
Freshness
Median age of returned records. Live lookups show as live.
Value
Cost per verified hit: price after misses and wrong answers are counted.

Badges

Verified
Identity, terms and data handling reviewed by Relaygrid.
Benchmarked
Ran this month against our labeled benchmark set of 2,000 records.
Grid Choice
Top Grid score in its capability, benchmarked, with 1,000+ calls. Providers that share ownership with Relaygrid are not eligible.

Scores come only from routed calls and the benchmark, never from providers. Providers can't pay for placement or edit scores.

The router is free. You pay for scale.

We give away the hard part: routing, one API, one key and MCP. Paid plans will charge for what grows with you: stored logs and records, queues and drains, and team governance.

Pricing details

Free forever

$0

The hard part, given away.

  • Routing and fallbacks

    Provider preferences, best-value routing and waterfall fallbacks on every call.

  • One API, one key

    One input and output schema per capability, whichever vendor answers.

  • MCP server

    Every operation is a tool your agent can call, with OAuth and scoped tokens.

  • Your own vendor keys

    Calls on your keys are never marked up or metered, on any plan.

  • Quality scores

    Fill rate, accuracy, latency and cost per verified hit, before you call.

Pay for scale

Not charged during launch

What we store and run for you. Free covers a lot of each.

Storage

Call logs, up to 360 days
Free plan: 100k calls a month, 30 days
Private cache records
Free plan: 50k records
Request and response bodies
Free plan: 1 GB

Past Free: $2 per 100,000 calls; $1 per 10,000 record-months (hot), $0.25 cold; $0.05 per GB-month

Operations

Saved waterfalls and pipelines
Free plan: 5
Queues, retries and rate-limit governance
Free plan: 1 queue
Log drains to your own tools
Free plan: Paid plans

Governance

Workspaces
Free plan: 2
Active API keys
Free plan: 25
Read-only SQL and anomaly alerts
Free plan: Paid plans

Security on every plan: envelope-encrypted key vault, hash-chained audit log, Postgres row-level isolation, scoped keys and agent tokens, vendor billing verification.

During launch everything is free and runs on your own vendor keys. None of the paid items is charged yet.

For data vendors

List your API in the catalog

Send us your API definition. We map it to a capability, review it and publish it. Scores come from routed calls, and no vendor can pay for placement.

Join 443 listed vendors

How listing worksEmail us to get listed
providers/acme-data/endpoints/email.yaml
- id: acme-data/email.find
  providerSlug: acme-data
  capabilityId: people.email.find
  name: Acme Data Email Finder
  description: Finds a work email from a name and company domain.
  input:
    first_name: { type: string, required: true, maxLength: 100 }
    last_name: { type: string, required: true, maxLength: 100 }
    domain: { type: string, format: domain, required: true }
  execution:
    mode: sync
    request:
      method: GET
      path: /v2/find
      query: { first: "{{input.first_name}}", last: "{{input.last_name}}",
               company_domain: "{{input.domain}}" }
  response:
    hit: { op: nonEmpty, path: result.email }
    map:
      email: { path: result.email, transforms: [lowercase] }
  pricing: { model: per_success, rate: { list: 6000 } }
  access: [managed, byok]

Keys stay in the vault

Security overview

  • Vault-encrypted keys

    Vendor keys are envelope-encrypted per workspace and only decrypted inside the call path. They never reach an agent's machine.

  • Scoped keys and tokens

    API keys and agent tokens are limited by capability, provider, spend cap and expiry.

  • Audit log

    Every call records who made it, with which key, to which provider, what it cost and the outcome.

  • Sealed call bodies

    Stored bodies are redacted, then encrypted with a per-workspace key. Viewing one is audited, and you can turn capture off.

Questions

How is this different from buying each tool directly?

One key and one schema per capability instead of a separate key and integration per provider. You also see measured fill rate and accuracy for every provider before you pick one.

What does GridRouter charge for?

Nothing during launch. Routing, the API, your key and the MCP server stay free. Paid plans will charge for what we store and run for you past the Free allowances: log history, cached records, request bodies, queues and team governance.

Where do the quality scores come from?

Only from calls routed through GridRouter over the trailing 30 days. Every score shows its sample size, and an endpoint without enough calls says it isn't measured yet instead of showing a number.

Do you mark up provider prices?

No. During launch every call runs on your own vendor keys, and calls on your own keys are never metered or marked up, on any plan.

What happens when a provider is down?

Failed calls aren't billed. A routed run falls through to the next provider on a miss or a failure, and the call object lists every attempt.

Can agents spend money without supervision?

Only up to their budget. Each agent has its own key with a daily, weekly, monthly or total cap, and calls stop when it is reached.

Is my CRM data used to train anything or shared with providers?

No. Payloads go only to the provider you call. Stored bodies are redacted and encrypted with a per-workspace key, you can turn capture off, and CRM writes run on your own OAuth grant.

One key. 443 vendors.

Free while we launch. Bring your own vendor keys; no card needed.

Get your free API keyBrowse the catalog