FreshBooks and Twilio integration

What moves between them

FreshBooks invoices, payments, bills, clients, and expenses trigger events that ml-connector captures via webhook or scheduled polling. For each event, ml-connector matches FreshBooks data to Twilio recipient lists (clients, staff email addresses, or phone numbers) and message templates (invoice created, overdue notice, payment received, bill approval request, vendor change alert). ml-connector then sends SMS or email notifications via Twilio, records the Twilio message ID, and polls Twilio's message status endpoint to capture delivery confirmation (sent, delivered, failed). If a message fails, ml-connector logs the error with full context so the notification can be re-sent manually or retried. Data flows one direction: FreshBooks events trigger Twilio messages, but Twilio message replies do not flow back into FreshBooks.

How ml-connector handles it

ml-connector stores FreshBooks OAuth credentials (refresh token) and Twilio API Key, refreshing the FreshBooks token when it expires and using Twilio HTTP Basic Auth on each request. It subscribes to FreshBooks webhooks for invoice, payment, bill, and client events, validates each webhook's HMAC-SHA256 signature against the registered secret, and parses the application/x-www-form-urlencoded payload to extract object_id and account_id. For each event, ml-connector looks up the corresponding FreshBooks object via REST (invoice details, client contact info, bill vendor email), maps it to message template placeholders (invoice number, amount due, due date, client name, vendor name), and resolves Twilio recipients by matching FreshBooks client email or phone to a configured recipient list. It then calls Twilio's Messages endpoint with the populated template, receives the message SID, and stores it in the audit trail. ml-connector polls Twilio's message status endpoint periodically to capture delivery status (delivered, failed, undelivered) and logs final status. If Twilio returns 429 (rate limit), ml-connector backs off and retries. If a message fails, ml-connector flags the record as failed-delivery so the event can be retried or escalated to manual follow-up. Because FreshBooks webhooks can arrive out of order or be delayed, ml-connector deduplicates by (object_id, account_id, event_type) to avoid sending duplicate notifications.

A real-world example

A mid-sized accounting firm handles invoicing and expense management for 50+ clients using FreshBooks. Before integration, the firm manually called or emailed clients when invoices were overdue and when bill vendors were added or changed, creating delays and inconsistent follow-up. Staff also missed payment confirmations and had to check FreshBooks daily for outstanding bills needing approval. With FreshBooks and Twilio connected, the firm now sends automatic SMS overdue reminders to client contacts 5 and 10 days past due, email confirmations to clients when payments arrive, and text alerts to approvers when new vendor bills arrive. The firm also sends vendor change notifications via email so clients know their invoices may be processed differently. Monthly manual outreach dropped by 70%, follow-up is instant and consistent, and the team spots payment issues before month-end close.

What you can do

• •Send SMS and email notifications from FreshBooks invoice, payment, bill, and client events to configured recipient lists.
• •Map FreshBooks invoices, amounts due, and due dates into Twilio message templates for overdue reminders, new invoice notices, and payment confirmations.
• •Authenticate FreshBooks via OAuth 2.0 user delegation and Twilio via HTTP Basic Auth with API Key, refreshing tokens as needed.
• •Validate FreshBooks webhook signatures (HMAC-SHA256) and Twilio message status callbacks (HMAC-SHA1) to ensure authenticity.
• •Track Twilio message delivery status (sent, delivered, failed) and log failed notifications for manual retry or escalation.

Questions

Can ml-connector send Twilio messages only to certain clients or vendors?

Yes. ml-connector resolves FreshBooks client email and phone number to a configurable recipient list per client or vendor, so you can target notifications to specific contacts. You can also use message templates with conditional logic to send different messages to clients versus internal staff based on the event type.

What happens if a FreshBooks webhook arrives late or a Twilio message fails to deliver?

ml-connector deduplicates FreshBooks events by (object_id, account_id, event_type) to prevent duplicate messages if a webhook is delayed or resent. If a Twilio message fails, ml-connector logs the failure with full context (invoice number, client, reason) so you can retry manually or configure an alert. Retries are tracked in the audit trail.

Does ml-connector support other notification channels besides SMS and email?

ml-connector currently supports Twilio SMS and email (via Twilio Messaging Services). Twilio also provides voice and video APIs, which can be integrated if your use case requires automated invoice-related voice calls or video notifications, but the standard integration focuses on SMS and email for invoicing workflows.