Meshes

Meshes is an embedded iPaaS for customer-facing SaaS applications, positioned as an "event-driven, developer-first embedded integration platform." The core problem it solves is this: when a SaaS product needs to deliver the product events it generates to the downstream tools each customer has connected (such as CRM, Webhooks, email tools, and Slack), developers often need to hand-write Webhooks, queues, workers, and retry logic for each destination, which makes multi-tenant routing, rate-limit handling, failure replay, and auditing fragmented and difficult to maintain. Meshes takes on this part of the "delivery mechanism" through a unified event model and routing layer.

Main capabilities

1. One event, multi-destination fan-out: the application only needs to emit one product event, and Meshes delivers it in parallel to all matching configured destinations (CRM, Webhook, and downstream systems).

2. Routing rules replace branching logic: define routing rules by event type and workspace, for example "when a contact is created on the website, send to HubSpot + Mailchimp," "when a lead is assigned and tier=Pro, notify Salesforce + a Webhook," "when a trial expires, only trigger an internal Webhook." Rules run per workspace, so each tenant or environment can have its own routing logic and limits.

3. Multi-tenant workspace isolation: each workspace has independent connections, rules, credentials, and limits, used to isolate customers, environments, or regions.

4. Credential management: handle OAuth, API keys, token refresh, and credential storage for each workspace, keeping tenant boundaries clear without hand-writing credential pipelines.

5. Retries, backoff, and dead letters: when a downstream API is rate-limited or fails, Meshes automatically retries, preserves failure context, and maintains a replayable path. Built-in exponential backoff handles rate limits.

6. Event history and visualization: record delivery status, logs, and metrics for each destination; search individual events to see where they went, what failed, why it failed, and what needs to be replayed.

7. Fast validation path: the quick start flow covers creating connections, creating rules, sending events, and verifying delivery, making it easy to validate the process end to end before connecting production traffic.

How to use it

In application code, only one event path is needed; send events via HTTPS (works with any service) or the SDK—the SDK is available but not required. Model events and resources once, and reuse them across connections and workspaces. At the connection level, authorize downstream tools and keep credentials scoped to thatTenant; at the rules level, define which events trigger which actions without writing routing logic into the application; at the delivery level, Meshes handles fan-out, rate-limit backoff, failure retries, dead-letter capture, and per-destination status, logs, and metrics. When downstream changes, routes can be modified in the UI or API without adding more branching logic in the application.

Use cases

• SaaS product teams: deliver the HubSpot, Salesforce, Mailchimp, Webhook, and other integrations customers expect without turning the application into an integration monolith; isolate customers, environments, or regions with workspaces while sharing the same core event model.
• Multi-tenant SaaS: isolate integrations for each customer without managing separate infrastructure, with each tenant having its own connections, rules, and event history, all managed in one view.
• Lead marketplaces: distribute a single lead to 2, 5, or 10 buyers at once, delivering via Webhook or direct CRM write, with retries, backoff, and a complete audit trail when downstream fails.
• AI agent builders: provide reliable external actions for AI agents; when an agent decides to send a Slack message, update a CRM, or trigger a workflow, Meshes handles delivery, retries, and observability.
• Growth and RevOps teams: connect new tools without making the engineering team add more queues or workers; developers define events, and rules and connections handle routing.
• Internal platforms: serve as the central event layer for internal tools, enrichment services, and data pipelines.

Limits and pricing (based on the materials)

The plans currently listed by Meshes are: the Developer plan includes 100 events per month, 1 workspace, 3 connections, 3 days of audit retention, and community support; higher tiers respectively include 2,500 events per month/8 workspaces/24 connections/7 days retention, 25,000 events/50 workspaces/250 connections/30 days retention and priority email support, 100,000 events/200 workspaces/2,000 connections/60 days retention, and custom commitments, pooled usage, longer retention, and launch support. The Developer plan can start for free. As for details such as the specific list of supported downstream destinations, SDK language coverage, and SLAs, a complete list is not yet provided in the materials, and the official documentation should be taken as authoritative.