Resources · n8n integration

n8n HTTP Request nodeConfigure HTTP Request in n8n.

Any REST API becomes reachable from a workflow. The n8n HTTP Request node sends a request to an endpoint you choose and hands the response to the next node. Its 18 parameters cover the method, the URL, authentication, and everything that keeps a long job alive.

Verified Trustpilot reviews · AI, automation & growth agency

Why automate

What does the n8n HTTP Request node actually do?

It turns a workflow step into an API call. You pick a method, type an endpoint, decide how the request authenticates, and the node returns whatever the server sends back as regular n8n data. It runs on every incoming item, so a list of ten records produces ten calls unless you group them. A trigger still has to open the workflow: this node answers a question, it never asks the first one.

The first job is reaching a service that has no node of its own. An internal billing API, a niche CRM, a public dataset: if it speaks HTTP, one node covers it, and the response lands in the same shape as any other node output. That is also the escape hatch when a dedicated node exists but stops short of the endpoint you need.

The second job is getting everything, not the first page. APIs cut long collections into pages, and Pagination walks through them until the response runs dry or a page cap is reached. A nightly export of orders into Google Sheets is a GET with pagination turned on, nothing more.

The third job is staying polite with a rate-limited service. Two hundred items firing at once earns a 429, so Batching splits them and puts a pause between groups, while Timeout stops a stalled call from holding the execution open. The node settings also carry Retry on Fail, with Max Tries and Wait Between Tries (ms), when the failure is temporary.

Prefer the dedicated node whenever one exists. Gmail and Slack ship their operations ready to fill in, with credentials already wired, and nobody gains from rebuilding that by hand. The comparison to check first is coverage: if the operation is in the dedicated node, use it.

The limits are honest ones. HTTP only: a file transfer over Ftp or a shell command needs its own node, and no URL field will bridge that gap. Self-hosted instances hit a second wall, since localhost inside a Docker container points at the container itself. Use http://host.docker.internal:5000, a Compose service name, or http://127.0.0.1:5000 instead. If you are still weighing platforms, the n8n review covers the hosting side.

Connect

How does the node authenticate a call?

  1. 01

    Start with None when the endpoint is open

    Leave Authentication on none and the node sends the request bare, with no token and no credential attached. Public datasets, health endpoints and internal services behind a private network work like that. Nothing is created in n8n, nothing is stored, and the panel stays short. The moment the server answers with a refusal instead of data, move to one of the two options below.

  2. 02

    Pick Predefined Credential Type when n8n already knows the service

    Set Authentication to predefinedCredentialType, then choose the service in Credential Type. n8n has already implemented the auth flow for many services, built-in and community alike, so an existing credential handles the signing. The official documentation recommends this route whenever the service appears in the list, because it also covers custom API operations without extra setup.

  3. 03

    Fall back to Generic Credential Type for everything else

    Choose genericCredentialType and the panel shows Generic Auth Type, where the method is configured by hand: Basic auth, Custom auth, Digest auth, Header auth, OAuth1 API, OAuth2 API, Query auth or Simplified Custom Auth. You supply the endpoints, the parameters and the method the API expects. A credential created once here is reusable across every workflow.

Parameters

What does each parameter change?

The HTTP Request node has 18 parameters. For each one: the node as you configure it in n8n, what the parameter changes, and our field notes.

01

Method

method

What you see in n8n

Notes & use cases

The dropdown at the top decides which verb travels on the wire, and the rest of the panel adapts to it. Reads use GET, writes use POST or PUT, partial updates use PATCH.

Key parameters

  • Method: seven choices, DELETE, GET, HEAD, OPTIONS, PATCH, POST and PUT, matching whatever the API documentation asks for.
Use cases
fetch a list with GET, then create a record with POST further down the workflow. HEAD earns its place when only the status matters and the body would be wasted traffic.
02

URL

url

What you see in n8n

Notes & use cases

Whatever endpoint you type here is where the call lands, and it is the one field the node refuses to run without. Expressions belong in it too, so the path changes per item.

Key parameters

  • URL: the full endpoint, shaped like http://example.com/index.html, often assembled from the previous node with {{ $json.id }} inside the path.
Use cases
walk a list of customer records and call /customers/{{ $json.id }} once per item. A typo or a retired route surfaces as the message about the resource not being found, so compare the path against the API documentation before blaming the node.
03

Authentication

authentication

What you see in n8n

Notes & use cases

Three ways to prove who is calling sit behind this selector, and the choice decides which extra field appears underneath.

Key parameters

  • None (none): the request goes out with nothing attached.
  • Predefined Credential Type (predefinedCredentialType): auth n8n already implemented, selected in Credential Type, recommended whenever the service is available.
  • Generic Credential Type (genericCredentialType): fully customizable, set in Generic Auth Type, from Basic auth to OAuth2 API.
Use cases
an in-house API guarded by a single token runs on Header auth in a couple of minutes.
04

SSL Certificates

provideSslCertificates

What you see in n8n

Notes & use cases

Some servers expect the caller to present a certificate of its own before the conversation starts. This switch is what makes the node offer one.

Key parameters

  • SSL Certificates: a plain toggle, left off unless the target insists on client certificates.
Use cases
a banking or health API behind mutual TLS rejects an anonymous client, and turning this on is what changes a refused handshake into an ordinary call. On a public API, where the server certificate alone does the job, the toggle stays untouched.
05

Send Query Parameters

sendQuery

What you see in n8n

Notes & use cases

Query parameters act as filters on the request. Turn this on when the API supports them and the call needs narrowing.

Key parameters

  • Specify Query Parameters: keypair (Using Fields Below) for name and value pairs added one by one, or json (Using JSON).
  • Query Parameters: the pairs themselves, extended with Add Query Parameter.
  • JSON: the same filters written as a single object.
Use cases
restrict an export to one status or one date range instead of downloading the whole collection and filtering afterwards.
06

Send Headers

sendHeaders

What you see in n8n

Notes & use cases

Headers carry metadata and context about the call, the kind of detail an API reads before it looks at anything else. This switch opens the fields that hold them.

Key parameters

  • Specify Headers: keypair (Using Fields Below) for name and value pairs, or json (Using JSON).
  • Headers: the pairs, added through Add Header.
  • JSON: the whole header set as one object.
Use cases
announce the format you accept, pass a version marker an API requires, or add a tracing header your own backend logs.
07

Send Body

sendBody

What you see in n8n

Notes & use cases

Creating or updating something means sending content along with the call. Switch this on, then match the format the API documents.

Key parameters

  • Body Content Type: json (JSON), form-urlencoded (Form Urlencoded), multipart-form-data (Form-Data), binaryData (n8n Binary File) or raw (Raw).
  • Specify Body: explicit fields (keypair) or a JavaScript object (json), filled through Body Parameters, JSON or a single Body such as field1=value1&field2=value2.
  • Input Data Field Name: the incoming field holding the binary file, paired with Content Type for raw payloads.
Use cases
push a signed PDF to a document service straight from the field that carries it.
08

Batching

options.batching

What you see in n8n

Notes & use cases

Input items become requests one for one, which a rate-limited API notices fast. Add this option to split them into groups and breathe between groups.

Key parameters

  • Items per Batch: how many input items go into each batch, with -1 disabling the split and 0 treated as one.
  • Batch Interval (ms): the pause between batches, 0 for none. Set 1000 to keep the pace at one request per second.
Use cases
the standard answer to a 429, alongside Retry on Fail in the node settings.
09

Ignore SSL Issues (Insecure)

options.allowUnauthorizedCerts

What you see in n8n

Notes & use cases

By default the node downloads a response only when certificate validation succeeds. Adding this option lets the response through even when validation fails.

Key parameters

  • Ignore SSL Issues (Insecure): a single switch, off in normal use, on when the certificate cannot be verified.
Use cases
a staging server with a self-signed certificate, reachable only inside the network, where the validation failure is expected rather than suspicious. On anything facing the public internet, fixing the certificate beats silencing the check.
10

Array Format in Query Parameters

options.queryParameterArrays

What you see in n8n

Notes & use cases

APIs disagree on how a repeated filter should look inside a URL, and this option picks the spelling, but only once Send Query Parameters is on.

Key parameters

  • No Brackets (repeat): the name repeats, foo=bar&foo=qux.
  • Brackets Only (brackets): square brackets after each name, foo[]=bar&foo[]=qux.
  • Brackets with Indices (indices): brackets plus a position, foo[0]=bar&foo[1]=qux.
Use cases
a 400 answer on a request carrying several values for one filter usually means the wrong spelling was chosen.
11

Lowercase Headers

options.lowercaseHeaders

What you see in n8n

Notes & use cases

Header names come out lowercase unless you say otherwise. Turned on is the default behavior, turned off keeps the capitalization exactly as typed.

Key parameters

  • Lowercase Headers: one switch, worth flipping only for a service that reads header names literally.
Use cases
an older internal API that compares a header name character by character will miss a value it should have found. Switching this off is faster than rewriting the service, and it changes nothing anywhere else in the request.
12

Redirects

options.redirect

What you see in n8n

Notes & use cases

When a server answers with a new location, the node follows by default. This option is where that behavior gets bounded or stopped.

Key parameters

  • Follow Redirects: on by default, off when you want to see the redirect answer itself rather than its destination.
  • Max Redirects: the ceiling on how many hops the request accepts before giving up.
Use cases
checking whether a shortened link still resolves means turning following off and reading the response. Auditing a chain of legacy URLs means keeping it on with a low ceiling.
13

Response

options.response

What you see in n8n

Notes & use cases

What comes out of the node is negotiable. This option shapes it, from the format to what counts as a failure.

Key parameters

  • Include Response Headers and Status: returns the full response instead of the body alone.
  • Never Error: succeeds even when the status code is not 2xx, so error handling moves into the workflow.
  • Response Format: autodetect (Autodetect), json (JSON), text (Text) or file (File), the last two writing into Put Output in Field.
Use cases
reading a rate-limit header requires the full response, not just the body.
14

Pagination

options.pagination

What you see in n8n

Notes & use cases

Large collections arrive in slices, and this option keeps asking until the collection is complete. Look at one unpaginated response first to see which shape the API uses.

Key parameters

  • Pagination Mode: off (Off), updateAParameterInEachRequest (Update a Parameter in Each Request) with its Parameters, or responseContainsNextURL (Response Contains Next URL) with Next URL.
  • Pagination Complete When: Response Is Empty, Receive Specific Status Code(s) with Status Code(s) when Complete, or Other with Complete Expression.
  • Limit Pages Fetched and Max Pages, plus Interval Between Requests (ms).
Use cases
a full customer export that no single call can return.
15

Proxy

options.proxy

What you see in n8n

Notes & use cases

Outbound traffic sometimes has to leave through a specific gateway, whether for filtering or for a fixed exit address. One field handles that.

Key parameters

  • Proxy: the HTTP proxy the request should use, written like http://myproxy:3128, taking precedence over the global HTTP_PROXY, HTTPS_PROXY and ALL_PROXY environment variables.
Use cases
a partner API that only accepts calls from an approved address, while the rest of the instance keeps its usual route. Because the field overrides the environment variables, the exception stays inside this one node.
16

Timeout

options.timeout

What you see in n8n

Notes & use cases

A server that never answers is worse than one that refuses. This value caps how long the node waits for response headers and the start of the body before it aborts.

Key parameters

  • Timeout: the wait in milliseconds, for instance 5000 on an endpoint that normally replies straight away.
Use cases
a scheduled workflow looping over hundreds of records cannot afford one stalled call blocking the whole run. Pairing a short timeout with Batching keeps a slow API from turning a nightly job into a morning problem.
17

Send Credentials on Cross-Origin Redirect

options.sendCredentialsOnCrossOriginRedirect

What you see in n8n

Notes & use cases

A redirect can point somewhere else entirely. This option decides whether the secrets follow it there.

Key parameters

  • Send Credentials on Cross-Origin Redirect: a switch that lets credentials, such as the Authorization header, travel on a redirect to a different origin.
Use cases
a download endpoint that hands the caller off to a storage domain still needs the header to accept the request. Outside that pattern, leaving the option out keeps a token from reaching a host that was never meant to read it.
18

Optimize Response

optimizeResponse

What you see in n8n

Notes & use cases

Attached to an AI agent as a tool, the node feeds its response to a model, and a raw API answer wastes tokens. This option trims it before the model reads it.

Key parameters

  • Expected Response Type: json (JSON) with Field Containing Data, Include Fields (All, Selected, Except) and Fields; html (HTML) with Selector (CSS), Return Only Content and Elements To Omit; or text (Text).
  • Truncate Response and Max Response Characters: a hard ceiling on the size.
Use cases
keeping only a handful of fields from a verbose catalog endpoint.
Need help

Need help automating HTTP Request with n8n?

A person reads every message.

FAQ

Questions people ask next

01Is the HTTP Request node free in n8n?
Yes. It is a core node, shipped with n8n, so there is nothing to install and nothing extra to pay on the n8n side. It behaves the same on n8n Cloud, the hosted offer run by n8n, and on a self-hosted instance installed through Docker or npm under the Community Edition and its Sustainable Use license. A workflow built in one place runs in the other. What the API on the far end charges is a separate matter, decided by that service and not by n8n. The node itself adds no metering, no per-call fee and no quota of its own.
02What do you need to make it work?
A URL, and whatever the target service expects in terms of authentication. There is no account to connect the way an integration has one: the node carries an Authentication selector with three settings. None sends the request bare. Predefined Credential Type reuses the auth n8n already implemented for a supported service, chosen in Credential Type, which the documentation recommends whenever it is offered. Generic Credential Type covers everything else through Generic Auth Type, with Basic auth, Custom auth, Digest auth, Header auth, OAuth1 API, OAuth2 API, Query auth and Simplified Custom Auth. A credential is created once in n8n and reused across workflows.
03What are the limits of the n8n HTTP Request node?
It speaks HTTP and nothing else. A file transfer over FTP, a shell command over SSH or a direct database connection each need their own node, and no amount of configuration in the URL field will change that. The Array Format in Query Parameters option only shows up once Send Query Parameters is on. Optimize Response is reserved for the case where the node is attached to an AI agent as a tool. Version 4 is the reference described here, and an older workflow may display an earlier version with fewer options, since a node never changes version on its own.
04When should you use HTTP Request instead of a dedicated node?
Use the dedicated node whenever it covers the operation. It lists the fields, validates them, and its credential is already understood by n8n, which is faster to build and easier for the next person to read. HTTP Request earns its place in two situations: the service has no node at all, or the node exists but stops short of the endpoint you need, which happens with newer or less common API routes. In that second case, Predefined Credential Type lets you reuse the service credential you already created, so the custom call authenticates without configuring anything by hand.
05n8n or Make for calling an API?
It comes down to hosting, data control and how the cost behaves. Make is a hosted platform with no self-hosting option, billed per operation, which is predictable when volume is stable and less so when a workflow loops over a large collection. n8n runs on n8n Cloud or on your own infrastructure through Docker or npm, so sensitive payloads can stay inside your network. Both build workflows visually. If the calls in question are numerous, paginated and carry data you would rather not send through a third party, self-hosting weighs in favor of n8n. Otherwise the choice is mostly about the team's habits.
Hack'celeration Lab

Get our weekly integration tips.

No spam. Unsubscribe anytime.