Harvest n8n integrationAutomate Harvest with n8n.

A new client record lands in Harvest and comes back with its ID, which the rest of the workflow reuses to attach contacts, projects or invoices.

• Account Name or ID: the Harvest account the client belongs to, picked from the list the credential loads.
• Name: the name of the client, the only content the node truly requires.
• Currency: sets the currency used for that client, which later estimates inherit when they leave it empty.
• Address: a textual address that accepts line breaks, useful for the block printed on documents.

The node sends a deletion for the client whose ID you pass, and the record leaves the account.

• Client ID: the ID of the client you want to delete, usually {{ $json.id }} from the node that listed or created it.
• Account Name or ID: the account holding that client, since an ID only makes sense inside one account.

One ID in, the full client record out: name, address, currency and active flag, ready to be read by the following nodes.

• Client ID: the ID of the client you are retrieving, not its name, which the API does not accept here.
• Account Name or ID: the Harvest account to read from.

Listing the client book is where Return All changes everything: enabled, n8n walks through the API pages and returns every client; disabled, it stops at the number set in Limit.

• Return All: on for a complete export, off when a preview of the first few clients is enough.
• Limit: the maximum number of clients returned when Return All stays off.
• Is Active: true keeps active clients only, false returns the archived ones.
• Updated Since: only clients modified after that date and time, which turns a full export into an incremental sync.

Changing a client that already exists means passing its ID and only the fields that move, since anything left empty in Update Fields stays as it was.

• Client ID: the ID of the client to update, the pivot of the whole operation.
• Is Active: false archives the client instead of deleting it, true brings it back.
• Name: the client name, for a rebrand or a legal entity change.
• Address: the postal address, including line breaks.

No ID to hunt for here. The operation returns the Harvest company behind the credential, which is the quickest way to confirm that the connection works and that it points at the right account.

• Account Name or ID: the only field to fill, and the one that decides which company is read when the login covers several accounts.

Contacts hang off a client in Harvest, so this operation always attaches the person to an existing company record.

• First Name: required, even when the rest of the identity comes later.
• Client ID: the ID of the client this contact belongs to, typically {{ $json.id }} from the client created just before.
• Email: the contact address, the field that matters for anything you plan to send.
• Title: the role of the person, handy when several people share one client.

When someone leaves a client's team, this removes the person from the account without touching the client itself.

• Contact ID: the ID of the contact you want to delete, which a list step usually provides.
• Account Name or ID: the account where that contact lives.

Pulling one contact back is the cheapest way to get an email address and a title without loading the whole list.

• Contact ID: the ID of the contact you are retrieving.
• Account Name or ID: the Harvest account to query.

The contact list comes back page by page, and the two filters decide how much of it the workflow has to read.

• Return All: enabled, every page is fetched; disabled, the run stops at Limit.
• Limit: the ceiling on returned contacts when Return All is off.
• Updated Since: restricts the answer to records changed after a date and time.
• Is Active: keeps active records, or returns the inactive ones when set to false.

Phone numbers, titles and even the client a person belongs to change more often than the contact itself, and Update Fields carries exactly what moved.

• Contact ID: identifies the record to modify.
• Client ID: moves the contact to another client, for a person who changed company.
• Email: the new address, in the name@email.com form the field expects.
• Phone Office and Phone Mobile: the two numbers stored separately, so updating one leaves the other alone.

An estimate starts life attached to a client, and Harvest numbers it on its own unless the workflow imposes a number.

• Client ID: the ID of the client this estimate belongs to, the only required content beyond the account.
• Subject: the estimate subject, the line the client reads first.
• Number: left empty, the number is generated automatically; filled, it follows your own numbering.
• Tax: a percentage applied to the subtotal, written as 10.0 for ten percent.
• Issue Date: defaults to today when the workflow leaves it out.

Draft clean-up is the usual reason to call this one: the estimate matching the ID disappears from the account.

• Estimate ID: the ID of the estimate you want to delete, which comes from a list step or from the creation that produced it.
• Account Name or ID: the account the estimate belongs to.

Before anything is sent to a customer, reading the estimate back gives its current state, amounts and subject as Harvest stores them.

• Estimate ID: the ID of the estimate you are retrieving.
• Account Name or ID: the Harvest account holding it.

Filtering by state is what makes this operation useful: the query can target only what is waiting for an answer.

• State: draft, sent, accepted or declined, one value per query.
• From and To: bound the period read.
• Client ID: narrows the answer to one client.
• Return All: on to page through everything, off to stop at Limit.
• Page: the page number in pagination, starting at 1, for a manual walk through a long list.

Amounts move between a first draft and a signature, and this operation rewrites only what the workflow provides.

• Invoice ID: the label the node uses for the ID of the document to update, so the value to pass is the estimate ID returned by the creation or the list.
• Discount: a percentage subtracted from the subtotal, 10.0 for ten percent.
• Notes: additional notes printed on the estimate.
• Purchase Order: the purchase order number the client asked to see.

Four fields are non negotiable here, because Harvest books an expense against a project, a category and a date.

• Project ID: the project the expense is charged to.
• Expense Category ID: the category it is tracked against, an ID rather than a label.
• Spent Date: the date the expense occurred, not the date the workflow runs.
• Total Cost: the total amount, or Units when the cost is computed from a quantity.
• Billable: true by default, so an internal cost has to set it to false explicitly.

Duplicate receipts are the common case: two imports of the same card statement, one line too many, and this operation removes it.

• Expense ID: the ID of the expense you want to delete.
• Account Name or ID: the account it was booked in.

Reading a single expense back returns its amount, project, category and billable flag, which is often what a validation step needs.

• Expense ID: the ID of the expense you are retrieving.
• Account Name or ID: the Harvest account to read.

Eight filters make this the widest query of the resource, and combining them is what keeps the run short.

• Is Billed: false lists what has not been invoiced yet, true what already has.
• Project ID and Client ID: restrict the answer to one project or one client.
• From and To: the period covered.
• User ID: the person who logged the expenses.
• Return All: on to collect every page, off to cap the result with Limit.

An expense is usually corrected after the fact, once someone notices the wrong project or a missing amount.

• Invoice ID: the label the node gives to the ID of the record to update, so pass the expense ID here.
• Project ID: rebooks the expense on another project.
• Expense Category ID: changes the category it is tracked against.
• Billable: flips a cost between billable and internal.
• Spent Date: fixes a date typed wrong at capture.

Billing starts here. The node opens a new invoice in Harvest for a client and returns the created record with its identifier, ready for the next node.

• Account Name or ID: the Harvest account to bill from, picked in the dropdown or passed as an ID with an expression, the n8n syntax that reads a field from the incoming item.
• Client ID: the client the invoice belongs to.
• Number: leave it empty in Additional Fields and Harvest generates the invoice number itself.
• Payment Term: upon receipt, net 15, net 30, net 45 or net 60.

Deletion is final on the Harvest side, so this operation usually sits at the end of a cleanup branch guarded by an IF node.

• Invoice ID: the invoice to remove, most often {{ $json.id }} coming from a Get Many step upstream.
• Account Name or ID: the account that holds the invoice, since an ID from one Harvest account will not resolve in another.

One invoice, one call. The node reads a single invoice and hands the full record to the rest of the workflow.

• Invoice ID: the invoice you are retrieving, typically stored in the CRM record or carried by the item that triggered the run.
• Account Name or ID: the Harvest account the lookup runs against.

This is the reporting workhorse of the Invoice resource. It lists invoices and pushes one item per invoice into the workflow.

• Return All: on, n8n walks the pages of the Harvest API until every invoice is in; off, Limit caps how many come back.
• State: a filter that keeps only draft, open, paid or closed invoices.
• From and To: bound the period; Updated Since narrows a run to what changed since the last execution.

An invoice that already exists can still move. This operation edits it in place and returns the updated version.

• Invoice ID: the invoice to modify. Everything else lives in Update Fields, so only what you fill gets touched.
• Due Date: pushes the deadline; left alone, Harvest keeps the date it computed from the issue date.
• Subject and Notes: the text a client reads on the document.
• Tax and Tax2: percentages applied to the subtotal, written as 10.0 for ten percent.

Time entries attach to projects, which makes this operation the first step of most client onboarding flows.

• Name and Client ID: what the project is called and who it belongs to.
• Is Billable: whether the hours logged on it can end up on an invoice.
• Bill By: none, People, Project or Tasks, the method Harvest uses to invoice the work.
• Budget By: the email of the user, or the literal value none.

Removing a project is the hard version of archiving it. Reach for it only when the project should leave Harvest for good.

• Project ID: the project to delete.
• Account Name or ID: the account it lives in.

Reading a project gives the workflow its settings, from the client it belongs to down to the budget fields, before anything else happens.

• Project ID: the project you are retrieving, often {{ $json.projectId }} taken from a time entry handled earlier.
• Account Name or ID: the Harvest account to query.

Pull the project list once and loop over it. Every project comes back as its own item.

• Return All: on, every project comes back page after page; off, Limit decides how many.
• Is Active: true keeps the live projects, false returns the archived ones.
• Client ID: restricts the list to one client.
• Updated Since: only projects touched after the date you pass.

Budgets and rates change during a client relationship. This operation applies those changes to an existing project.

• Project ID: the project to update; fields left empty in Update Fields stay as they are.
• Budget and Budget Is Monthly: the hours allowed, and whether that allowance resets every month.
• Hourly Rate: used when the project is billed by project hourly rate.
• Is Active: set it to false to archive the project instead of deleting it.

Tasks are the activity labels a person picks when logging time, such as design or support. This operation adds one to the account.

• Name: the only required field beyond the account, so the node stays short.
• Default Hourly Rate: the rate applied when the task is added to a project, 0 unless you set it.
• Is Default: on, the task is added automatically to future projects.
• Billable By Default: whether the hours logged on it count as billable.

Harvest drops the task from the account when this runs, which also removes it from the picker people use when tracking time.

• Task ID: the task you want to delete.
• Account Name or ID: the account it belongs to.

Use it to confirm what a task ID actually points to before another node writes anything.

• Task ID: the task you are retrieving.
• Account Name or ID: the account to read from.

The task list is short in most accounts, which makes it a good source for dropdowns and lookups built elsewhere.

• Return All: on, the node collects the whole list; off, Limit stops it earlier.
• Is Active: true for the tasks people can still pick, false for the archived ones.
• Page: moves through the pages by hand when Return All stays off.

Renaming or repricing a task keeps every past time entry intact, unlike a delete.

• Task ID: the task to update.
• Name: the new label, visible immediately to the people tracking time.
• Default Hourly Rate: the rate future projects will inherit.
• Is Active: false archives the task without erasing its history.

When the workflow already knows how long something took, this is the operation that writes it to Harvest.

• Project ID and Task ID: what the hours are charged to. Both are required, so keep a lookup in front of this node.
• Spent Date: the date the time was spent, in ISO 8601 format.
• Hours: the amount of time tracked.
• User ID: defaults to the authenticated user, so set it when the entry belongs to someone else.

Same result, different input. Here Harvest gets two clock times and works out the duration itself.

• Started Time and Ended Time: clock values such as 8:00am and 3:00pm. Left empty, the start defaults to the current time.
• Spent Date: the ISO 8601 date the entry belongs to, which can differ from the day the workflow runs.
• Notes: free text that explains the entry to whoever reviews it.

Hours logged on the wrong project are worse than no hours at all. This operation takes the entry out of Harvest.

• Time Entry ID: the entry you are deleting.
• Account Name or ID: the account it was logged in.

Narrower than it first looks. The entry stays in Harvest with its hours and notes; only the link to the outside record it came from is cut.

• Time Entry ID: the entry whose external reference you are deleting.
• Account Name or ID: the account holding the entry.

Reading a single entry gives the workflow the detail behind a line in a report.

• Time Entry ID: the entry you are retrieving.
• Account Name or ID: the account to read from.

Everything invoicing and reporting needs comes out of this operation. Filters decide how much of the history comes with it.

• Return All: on, n8n pages through the full history, which can be long on an old account; off, Limit keeps the run small.
• From and To: the period the entries were spent in.
• Is Billed: false isolates the hours not yet invoiced.
• Is Running: true returns only timers still counting.
• User ID and Client ID: narrow the list to one person or one client.

Harvest picks the timer back up on an entry that was stopped earlier, instead of creating a second one for the same work.

• Time Entry ID: the stopped entry to restart. Harvest refuses the call if that entry is already running, so branch on the state before you send it.
• Account Name or ID: the account the timer lives in.

The opposite move, and the one that saves the most hours at the end of a month. It closes a timer that is still counting.

• Time Entry ID: the running entry to stop. The call only works while it is running, so pair it with a Get Many filtered on Is Running.
• Account Name or ID: the Harvest account concerned.

Corrections belong here rather than in a delete followed by a fresh entry, because the entry keeps its place in the timeline.

• Time Entry ID: the entry to fix.
• Hours: the corrected amount of time tracked.
• Started Time and Ended Time: the clock values, when the entry was created from a start and end pair.
• Notes: the place to record why the line changed.

Adds a person to the Harvest account and returns the new user record, identifier included.

• Account Name or ID: the Harvest account the user lands in, picked from the dropdown or passed through an expression, meaning a value borrowed from a previous node, such as {{ $json.accountId }}.
• First Name, Last Name and Email: the three identity fields Harvest asks for, the email box showing name@email.com as its example format.
• Additional Fields: 13 optional settings, among them Is Admin, Default Hourly Rate and Weekly Capacity, counted in seconds and set to 126000 by default, which is 35 hours a week.

Deletion here is final, so this operation belongs at the end of an offboarding run, once everything that had to be billed has been billed.

• Account Name or ID: tells the node which Harvest account holds the person, from the list or from an expression.
• User ID: the ID of the user you want to delete. Harvest expects that identifier, not an email address or a display name, so read it from an earlier step with {{ $json.id }}.

Fetches one user record from its identifier and passes the full profile down the workflow, which saves a manual lookup when only an ID has been travelling between nodes.

• Account Name or ID: the Harvest account to query, selected in the node or resolved by expression.
• User ID: the ID of the user you are retrieving, most often {{ $json.id }} when the value comes from a list step earlier in the run.

Returns the people attached to an account, one item per user, which is the usual starting point for anything that loops over a team.

• Account Name or ID: the account whose roster gets listed.
• Return All: enabled, n8n follows every page of the Harvest API until the roster is complete; disabled, Limit caps how many users come back.
• Filters: 3 options, Is Active to separate current people from archived ones, Updated Since to keep only records touched after a date, and Page to target a specific page of results.

Reads back the profile tied to the credential itself, which is the shortest way to confirm that the token or the OAuth2 connection really reaches Harvest.

• Account Name or ID: the only field to fill, and it doubles as a check that the credential can actually see that account.

Changes what Harvest stores about someone already in the account, so rates, permissions and capacity move without creating a second record for the same person.

• Account Name or ID: the account holding the record to edit.
• Time Entry ID: the node labels the required identifier this way for this operation, and its description reads "The ID of the time entry to update".
• Update Fields: 16 editable settings, including Email, First Name, Last Name, Is Active, Is Project Manager and Cost Rate.