Can you automate workflows in Fibery?

Quick Answer: Yes. Fibery ships two native automation types, Rules that fire on entity creation, update, linking, unlinking or a schedule, and Buttons that a user clicks, and both are available on the free plan within a 250-run monthly allowance. No-code Rules already carry filters and run their action steps in sequence; JavaScript scripted actions, which add arbitrary logic and outbound calls with branching, require the Pro plan or above, and external orchestration is available through an HTTP commands API, a per-Space GraphQL endpoint, outbound webhooks and a hosted MCP server.

The two native automation types

Fibery calls the feature set Automations, and it has exactly two members. Rules run when a trigger condition is met on a Database: an Entity is created, updated, linked to another Entity, unlinked, or a schedule fires. Buttons run the same kinds of action when a user clicks them, in the context of the clicking user and using that user's authorisations to external services. Filters narrow which entities a Rule acts on, and action steps always execute in sequence.

Available actions include create, update and delete an Entity, append content to a rich text field, add a comment, add or clear assignees, move to a final state, attach a file from a URL, notify, collection operations such as add, link and unlink, and native steps for GitHub, GitLab, Jira, Slack and Email. Validation rules, which block a save rather than react to one, are a separate feature restricted to Pro and Enterprise.

flowchart LR
  T{Trigger} -->|Created / Updated / Linked / Schedule| F[Filter]
  B[Button click] --> F
  F --> A1[Native actions]
  F --> A2["Script (Pro+)"]
  F --> A3[Send Web Request]
  A3 --> X[(External service)]
  A2 --> API[Fibery API / GraphQL]

The limit that changes the maths

As of August 2026, monthly run allowances are 250 on Free, 1,000 plus 100 per paid seat on Standard, 10,000 plus 1,000 per paid seat on Pro, and unlimited on Enterprise. Those numbers look small until batching is accounted for: one run processes up to 50 entities. A Button clicked once across 100 selected entities costs two runs, not 100. Runs are only counted when a Rule actually executes, so entities filtered out are free, and counters reset on the calendar month rather than the billing cycle. Extra usage, the metered overage option, exists only on Pro and Enterprise. On Free and Standard, exceeding the allowance stops Rules from running until the next calendar month or an upgrade, with no way to buy through it. Admins are notified at 75% of the limit if that is reached before the 20th of the month and at 90% if reached before the 25th.

Where the paywall sits

The consequential gate for an automation audience is scripting. JavaScript scripted actions require Pro or Enterprise, and a Rule containing a script can only be configured by an Admin, although any Architect can build a scripted Button. Scripts get three services: fibery for entity and schema operations including raw commands and in-script GraphQL, http for outbound calls, and utils. All Fibery calls are asynchronous.

The no-code alternative for outbound calls is Send Web Request, available on lower plans. It supports GET, POST, PUT, PATCH and DELETE with templated URLs and headers, and imports from cURL. Two constraints matter: it is HTTPS-only with private IP ranges and the workspace's own hostnames blocked, and it is fire-and-forget. The run history shows the response status and a truncated body, but a later action step cannot branch on the result.

External orchestration

For anything Fibery's own engine cannot do, the platform exposes an HTTP commands API at POST /api/commands, per-Space GraphQL endpoints, and outbound webhooks configured per Database. Rate limits, as published in August 2026, are 3 requests per second per token and 7 per workspace. Two gaps are worth knowing before designing around them: webhooks do not fire on rich text field changes, and Fibery has no outbound concurrency control, so a batched run lands at the receiver almost simultaneously and throttling has to happen at the receiving end.

Connector support across the major orchestrators, verified August 2026: Zapier has 7 operations and polls rather than firing instantly; Make has 6 modules with instant triggers; n8n has no built-in node but a verified community node built and maintained by Fibery; and Pipedream has an app with 2 triggers and 7 actions, authored by Pipedream rather than Fibery. Zapier, Make, the public API and GraphQL are all available on the Free plan.

Agents

Fibery runs an official hosted Model Context Protocol server at mcp.fibery.io with OAuth authentication, listed as unlimited on all four plans including Free. It covers querying, search, entity and document writes, workflow state, and live schema modification, which means an external assistant can create Spaces, Databases and Fields rather than only reading data. Fibery has stated publicly that it will not invest heavily in its own internal agents and is betting on external agents through MCP instead. One capability regressed: the ability for Fibery AI to create Rules and Buttons was removed on 2 April 2026 and, as of the July 2026 product report, was still being rebuilt.

Editor's Note: Counting the connected tool roster on 18 August 2026, Fibery's hosted MCP server exposed 80 tools against the 50 listed in its developer tool-reference table, and its built-in connector list returned 28 connectors of which 26 support continuous sync. Both numbers are higher than any figure Fibery publishes in marketing. The caveat cuts the other way for orchestration: 27 of those 28 connectors are one-way into Fibery, and Jira is the only two-way sync, limited to Name, Priority, Status, Assignee and a few custom field types. If your design assumes Fibery writes back to the source system, check that assumption first. — Rafal Fila, ShadowGen

Written & reviewed by Rafal Fila · Last updated:

Related Tools