Articles in this section

Current release: 2026.8.1

Michelle Welch
Updated

The August release adds Celigo Storage, a platform-managed file storage resource usable directly inside flows, APIs, and tools. Enterprise MCP gains end-user access controls, capability sets, and JIT provisioning through an external IdP. Celigo Sync adds Google BigQuery as a destination. Connectivity expands with a Google Cloud Storage connector, IAM role authentication for Amazon Redshift and AWS HTTP connections, and execution logs for the Salesforce connector. The release also brings API execution logs with trace view and pagination in API Builder, unified EDI trading partner configuration in B2B Manager, and schema validation, lookup cache, and mapping improvements across the platform. 

AI automation

  • Enterprise MCP — end-user access and IdP provisioning: This release delivers a complete enterprise access model for MCP servers. End users are a new lightweight user type for people who consume MCP tools without needing integrator.io access. You can organize them into groups, map groups to MCP servers with specific tools enabled, and define capability sets to bundle related capabilities.  JIT provisioning auto-creates end-user accounts on first login via an external IdP, with group memberships synced automatically from OIDC group claims on each login. Multiple IdPs are supported for end-user authentication, with email domain routing separate from account SSO. To learn more, see Create and manage end user groups, Configure end user authentication, Invite and manage end users, and Configure group sync for external IdPs.
  • AI Agent Chat — interactive development and testing: AI Agents now include a built-in chat interface for real-time testing inside the configuration screen, without saving first. Tool calls stream live status, failures surface immediately with context, and a Debug Mode provides expandable trace sections for tool, prompt, and resource execution. Conversation history is saved per agent and filterable by recency. Learn more about conversational agents.
  • MCP server request log — user and access token visibility: Two new columns in the MCP server request log — User and Access Token — attribute every tool call to a specific person and the credential used. Supports per-user auditing and consumption tracking for SOC 2 and other compliance requirements. See Manage MCP server.
  • One-to-many support for AI agent and guardrail imports: One-to-many is now available on AI agent and guardrail imports, so a single source record can generate multiple destination records from a child array — matching the behavior already supported on other import types. Learn more about One-to-many support for AI agent and guardrail imports

Connectivity

  • Celigo Storage: A new Storage resource is available under Resources in integrator.io, giving you a platform-managed place to store, organize, and browse files without external SFTP servers or S3 buckets. Folder management and role-based access control are included. Celigo Storage is also usable directly inside flows, APIs, and tools as a built-in file provider — no external file system required. File-based export, import, and per-record lookup steps are available on the Celigo APIs connector, alongside record-based admin operations (list, delete, create folder, move, rename). Celigo APIs connections can be scoped to specific integrations so builders without account-wide admin access can use Celigo Storage in their flows, and admins can restrict a Custom-scoped connection to specific folder paths, limiting read and write access to only the folders a connection needs. Learn more about this feature.
  • Google Cloud Storage connector: Integrator.io now features a built-in Google Cloud Storage connector using Google service accounts (simply paste your JSON key once). Bucket names automatically populate from your GCP project, complete with a browse button for directory paths. Built on the HTTP 2.0 framework, it supports both file transfer (export, import, lookup) and record-based administrative operations (list, get, delete, copy objects; list/get bucket metadata). Users can switch between a Simple view and an HTTP view for direct access to any GCS JSON API endpoint.
  • IAM role authentication for Amazon Redshift: Redshift connections can now use a cross-account AWS IAM role with temporary, auto-rotated credentials instead of long-lived database username and password. The same AWS IAM iClient used for S3 can be reused, and a new Help Center article provides ready-to-paste IAM policy and trust-relationship JSON for AWS Console.
  • IAM role authentication for AWS HTTP connectors: AWS Service (Generic), AWS Textract, and AWS Bedrock connections now support IAM role-based authentication with temporary, auto-rotating credentials. If you've already configured an AWS IAM iClient for S3 or Redshift, you can reuse it here without additional setup. 
  • Execution logs for the Salesforce connector: Salesforce flows now generate per-record execution logs in production across three levels (Basic, Detailed, Debug)—logging endpoints, status codes, methods, and direct record URIs for imports (SOAP/REST/Composite REST), exports, and lookups (REST), with Detailed/Debug levels adding full payloads and headers. 
  • HTTP redirect handling controls: HTTP exports and imports now include an Automatically Follow redirects option (on by default) and a configurable redirect cap (1–10). When this option is off, the 3xx response — including its status code and Location header — is returned to your response mapping. A new connection-level option removes the Authorization header on cross-host redirects.
  • Salesforce Data Cloud connection support: The HTTP connector now supports Salesforce Data Cloud's Ingestion API, using a standard Salesforce Connected App and JWT-based authentication. Celigo handles the full token exchange and refresh automatically.

Celigo platform

  • Lookup cache access via Handlebars: You can now read Lookup cache values in Handlebars expressions across AFE-enabled dynamic fields using the Lookup cache alias, eliminating the need to duplicate lookup logic outside of mappings. Learn more about the getDataLookupCache handlebars helper.
  • Advanced schema validation in the UI: The schema editor now supports enum, regex, numeric range, length, and array-size validations in Manual mode, with a Manual↔JSON toggle. Previously, any validation beyond basic required/optional checks required hand-editing JSON. Switching from JSON to Manual mode now shows a warning before any unsupported settings are dropped. Learn more about Advanced schema validation. You can also refer to Configure tool input and Configure tool output.
  • Enhanced dateAdd handlebars helper: The dateAdd handlebars helper now accepts an optional interval unit (milliseconds, seconds, minutes, hours, days, weeks, months, or years), so you can add or subtract time using readable units instead of converting to milliseconds; templates that omit the unit continue to work unchanged.
  • Lookup cache usage tracking: Your account's Lookup cache consumption is now visible from your subscription, with a per-cache usage breakdown, up to 14 months of usage history, and a direct path to request an upgrade when nearing your plan's limit. Learn more about lookup cache usage tracking.
  • Tree-based field selector for Branching and Filters: Branching and Filters now use the same searchable tree-based field selector as the mapper. Nested source fields are browsable with inline data type display, replacing the flat dropdown that made deep or large payloads difficult to navigate.
  • Hardcoded values for Object and ObjectArray types in Mapper 2.0, Transformations 2.0, and body params: Mapper 2.0, Transformations 2.0, and body params now accept hardcoded values for Object and ObjectArray destination fields, not just JSONPath or Handlebars expressions. This covers exports, lookups, and imports, and applies wherever only expression-based mapping was previously allowed for these types.

Data ingestion

  • Google BigQuery as a destination in Celigo Sync: Google BigQuery is now a supported destination in Celigo Sync, allowing you to ingest data from 600+ sources into BigQuery with automated schema management. Historical backfills and incremental syncs are handled automatically.

Celigo headless

  • OAuth sign-in for the Celigo Platform MCP server: AI assistants like Claude and Cursor can now connect to Celigo with a browser sign-in — no API tokens to generate or paste. Add Celigo from an AI assistant marketplace with a Connect button; your existing Celigo email, SSO, and MFA all work. Sign-in is scoped to one account and one environment, and connections stay signed in automatically via OAuth 2.1, with unused connections revoked after a period of inactivity. Existing API token and Personal Access Token connections continue to work unchanged.
  • Credential-free PATCH for connection concurrency and rate-limit settings: Connection concurrency level, borrowed concurrency, and rate-limit auto-recovery settings can now be updated via the integrator.io API without re-supplying credentials. Useful for programmatic throughput tuning from the API, CLI, or an MCP server.
  • Audit log source values for Ora, CLI, and MCP: The audit log now distinguishes changes made through Celigo Ora, the Celigo CLI, or the Platform MCP server as distinct Source values, alongside the existing UI and API sources. All operation types — creates, modifications, and deletions — are attributed, and the new values are filterable.

B2B Manager

  • Unified EDI trading partner connector configuration: Protocol (AS2, VAN, FTP, and others), business region, and business model are now configurable within a single adaptive connector, replacing a catalog of discrete protocol-specific connector variants. Connector behavior adapts dynamically to the selected combination, and only compatible options are shown based on the selected trading partner and business model. Learn how to set up a connection to a trading partner.

API Management

  • API pagination in API Builder: APIs built in API Builder can now return paginated responses. When a result set exceeds the configured page size, the response includes a pagination object with has_more and a celigo_next_cursor for retrieving subsequent pages. Supported when the first lookup in the API uses an HTTP, NetSuite, or Salesforce adapter.
  • API execution logs with trace view: API Builder now includes end-to-end execution logs with step-level trace views, covering request and response steps, lookups, imports, mappings, transformations, and branching in a single view. Five logging levels —No Logging, Basic, Standard,  Detailed, and Debug — let you dial up the details when needed.

Integration apps & templates

  • Shopify Payout sync to NetSuite Deposits (GraphQL migration & configurable account mapping): Shopify Payments payouts and their underlying transactions (sales, refunds, fees, adjustments, Shop Cash, Shop Pay Installments, and Shopify Collective orders) now sync to NetSuite as bank deposits using GraphQL instead of REST, with a new configurable mapping table that routes each transaction type, including chargebacks, to the correct GL account. For more information, see Configure the GraphQL payout flow.

Celigo CLI

  • @celigo/celigo-cli 2026.8.2 is on npm and includes everything in 2026.8.1. The list verb now projects fields server-side by default — one account's exports list dropped from 12.9 MB to 183 KB — and new storage, datasets, and processors command groups join full create, update, and delete support in syncs and background installation of the official Celigo skills. This release also includes breaking changes to date and time option names, flows run window flags, and the removal of --sandbox from three commands. See the developer CLI change log.