Celigo November 2026 Release

MCP Management

Make MCP tools aware of who’s calling them

MCP tools can now access information about the authenticated caller to determine what information to retrieve or how to process a request.

Rather than building separate tools for each employee, teams can use caller context to create shared tools that return results appropriate to each user.

Caller context lets you:

• Reference the caller’s email, name, and user ID in queries, mappings, filters, headers, and scripts.
• Filter results using the caller’s groups and roles when supported by the authentication configuration.
• Attribute actions performed through AI assistants to the user who initiated them.
• Identify the credential and MCP server associated with a request.

For example, a sales representative asking an AI assistant for open opportunities can receive only the opportunities assigned to them. A finance employee using a different tool can retrieve records by department or role.

Celigo resolves the caller’s identity from the authenticated MCP request rather than accepting identity information supplied by the AI model. When a call has no associated person, caller-dependent filters can return no matches rather than unrestricted results, provided the tool is configured to preserve those filters.

Learn more about caller context in MCP tools

Discover MCP servers through a central registry

The MCP server registry API provides a central endpoint to discover the MCP servers that a user or application is authorized to access.

Instead of manually maintaining server URLs across development environments and applications, teams can dynamically retrieve available servers from the registry.

The registry lets you:

• List MCP servers based on the caller’s existing permissions.
• Retrieve server connection information.
• Reflect newly available servers and changes in access.
• Support compatible scripts, developer tools, and internal catalogs through the MCP Registry v0.1 specification.

Discover MCP servers with the MCP server registry

Manage MCP access across teams with roles

End-user roles simplify MCP access management by combining permissions for multiple MCP servers into a reusable role that can be assigned to individuals or groups.

Roles let administrators:

• Grant access to several MCP servers and their capability sets in one action.
• Update permissions for everyone assigned to a role by editing it once.
• Review each end user’s effective access, including permissions inherited through roles, groups, or direct assignments.

Roles are supported on MCP servers configured with Celigo Identity. Existing direct and group grants continue to work, and administrators can reset an end user’s MFA without removing their access.

Learn how to create and manage end-user roles

Additional MCP Management enhancements

Builders can deploy MCP servers more easily, investigate failures with greater visibility, and manage production and sandbox environments independently.

Upload MCP servers and richer Marketplace templates

Move MCP server configurations between accounts and environments without rebuilding them. Packages include capability sets, resources, and upstream connections. Credentials and access grants don’t transfer, and installed servers remain disabled until reviewed.

Find and install an MCP server template

MCP server logging and retry job visibility

Troubleshoot MCP calls with step-level execution logs showing connections, mappings, requests, responses, and errors. Retry jobs now capture execution details, making it easier to understand why processing failed.

View MCP server request history and execution logs

Per-environment Platform MCP URLs

Connect production and sandbox environments independently in supported MCP clients. Dedicated URLs keep authentication scoped to each environment, allowing builders to work across environments without repeatedly switching connections.

Agentic automation

Knowledge bases for AI agents

AI agents need business-specific context to deliver useful answers. Knowledge bases make company information accessible during agent execution, helping reduce reliance on general model knowledge and enabling more relevant, context-aware responses.

Key capabilities include:

• Ground agents in company knowledge: Create knowledge bases in AI Studio using documents stored in Celigo Storage. Attach them to AI agents so they can retrieve relevant information from company policies, procedures, and other approved documents when responding to requests.
• Keep knowledge current: Celigo automatically indexes documents and reindexes them as content is updated. Configure freshness rules to determine how agents handle outdated information, and use execution logs to see which knowledge bases and files were retrieved.
• Reuse knowledge across agents: Share a single knowledge base among multiple AI agents, allowing different teams to reference the same business information without maintaining separate document collections.

See create and manage knowledge bases.

Code execution for BYOK AI agents

AI agents using supported bring-your-own-key model connections can now execute code as part of their tasks.

Add Code Execution from the agent’s Add tools panel to perform calculations, transform data, or generate files without creating a separate script step.

The capability initially supports selected OpenAI, Gemini, and Anthropic models, with Anthropic support in beta. Code runs in the model provider’s sandbox, execution details appear in the agent trace, and supported output files can be used in subsequent flow steps.

Model lifecycle status in the agent model picker

Identify AI agents and guardrails that use models approaching retirement.

The model picker now shows lifecycle statuses, including supported, retiring, retired, and unavailable. Warnings identify affected agents and guardrails, while suggested replacements help builders plan model updates.

Retired or unavailable models can block saves and execution, with an explanation of the issue. Celigo does not automatically switch an agent to another model.

Understand AI model lifecycle labels

Platform

Track file activity across Celigo Storage

Celigo Storage now provides a dedicated activity log showing who accessed or modified files, what action they performed, and when it occurred.

The log captures file operations from the Storage interface, APIs, integration flows, AI agents, the CLI, and SFTP file server accounts.

You can:

• Track uploads, downloads, renames, moves, copies, and deletions.
• Identify the user or trading partner account responsible for an operation, along with available source and IP information.
• Filter activity by action, user, source, or date.
• Download filtered activity records as CSV for investigations and security reviews.

Activity logging starts automatically and follows existing Celigo Storage permissions, so users see activity only for folders they are authorized to access.

Learn more about the Celigo Storage activity log

Give SFTP partners only the access they need

Celigo-hosted SFTP now supports more precise permissions and additional authentication options for trading partner accounts.

You can:

• Configure independent Read, Create, Modify, and Delete permissions.
• Create write-only drop boxes for inbound partner files and read-only folders for outbound documents.
• Authenticate partners using supported SSH public keys, including Ed25519, ECDSA, and RSA 2048-bit or larger.
• Rotate SSH keys without interrupting transfers by keeping multiple keys active.
• Allow users with Manage access to an integration to administer file server accounts associated with its folders.

Existing accounts retain their current access until permissions are updated. Permission and credential changes are recorded in the audit log.

Manage SSH keys for a file server account

Replay events and troubleshoot flows with Celigo Topics

Celigo Topics is adding more ways to recover from processing issues, investigate event activity, and move event-driven integrations across environments.

Replay events without interrupting live processing

If an outage, mapping error, or downstream problem affects a flow, you can replay previously consumed events from the earliest retained message or a specific time window. Replay runs as an independent job for the selected flow, without changing the live subscription cursor.

Investigate Topics activity in execution logs

Topic publishes and deliveries appear in execution logs, including topic names, timestamps, and message offsets. Detailed and Debug logging can include payload information.

Promote Topics between environments

Cloning, template exports, environment promotion, and integration lifecycle management carry Topics configuration into the target environment.

For example, if an incorrect mapping affected orders during a particular afternoon, you can fix the mapping and replay the affected window while new orders continue to be processed.

Messages and subscription positions do not transfer between environments. Replayed messages may also be delivered more than once, so downstream processing should tolerate duplicates.

Celigo Topics overview

Additional platform improvements

Execution logs for tools in flows and APIs

Trace tool execution within a flow or API, including lookups, imports, mappings, hooks, and nested tools up to five levels deep. Execution details follow the parent flow’s or API’s logging level, permissions, and retention settings, helping builders identify where processing failed without having to reproduce the entire run.

View flow run logs

Celigo File storage connector

Celigo Storage now has a dedicated connector in the application catalog. Builders can transfer and look up files, manage folders, and perform supported file operations without navigating through the Celigo APIs connector. Existing Celigo APIs storage connections and steps continue to work.

Set up a connection to Celigo file storage

Automatic page splitting after lookups

When lookup enrichment causes a flow page to exceed 5 MB, Celigo can automatically split the page between complete records and continue processing. This reduces manual page-size adjustments and flow failures caused by growing lookup responses. Individual records exceeding 5 MB remain subject to the limit, and the behavior does not yet apply to test runs.

Network policies for webhook listeners

Restrict which networks can invoke webhook listeners using reusable network policies. Requests from disallowed addresses receive HTTP 403 before the flow starts. Existing listeners remain unrestricted until a policy is assigned.

Create webhook listeners

Improved multi-environment upgrades

Accounts moving from legacy production and sandbox configurations to multiple environments receive clearer readiness checks, upgrade progress information, and recovery options if an upgrade is interrupted.

Data ingestion

Replicate operational data into Amazon Redshift

Celigo Data Ingestion now supports Amazon Redshift as a managed sync destination alongside Snowflake, Google BigQuery, NetSuite Analytics Warehouse, and Microsoft SQL Server.

You can:

• Replicate operational data from Salesforce, NetSuite saved searches, and other supported sources into Redshift.
• Automatically create destination schemas and tables during the initial sync.
• Apply supported source schema changes during subsequent runs.
• Load data using Redshift’s native COPY command with Celigo-managed staging.
• Configure merge, append, or replace behavior for destination objects.

For example, a revenue operations team can replicate Salesforce accounts and opportunities into Redshift for reporting. As supported fields change in Salesforce, Celigo updates the destination schema without requiring manually maintained warehouse scripts.

Because Celigo manages the staging process, customers don’t need to provision an additional S3 bucket or configure a separate COPY pipeline.

Existing Redshift imports and flows continue operating as configured. To use managed Data Ingestion, create a new sync and select Amazon Redshift as the destination.

Additional Data Ingestion capabilities planned for November

Azure Synapse Analytics as a sync destination

Synchronize operational data into Azure Synapse dedicated SQL pools with managed schema creation, schema updates, and optimized bulk loading through COPY INTO. Data staging takes place in the customer’s Azure Data Lake Storage Gen2 container.

Microsoft Fabric Warehouse as a sync destination

Replicate supported operational data into Microsoft Fabric Warehouse with managed schema creation and loading, using staging within the customer’s Fabric environment.

Trigger flows after Data Ingestion sync completion

Configure a Data Ingestion sync to trigger one or more integration flows after completing a run. This supports downstream processing such as warehouse transformations without relying on a fixed schedule that may run before data loading finishes.

B2B Manager

See the complete EDI order lifecycle in one place

The EDI lifecycle view connects related documents across an order’s lifecycle, helping teams understand where a transaction stands without manually matching individual EDI records.

The updated dashboard provides:

• Lifecycle tracking: View related purchase orders, acknowledgments, advance ship notices, and invoices together.
• Expanded search: Find transactions using purchase order numbers, control numbers, and other supported document identifiers.
• Acknowledgment visibility: Use color-coded statuses to identify accepted, rejected, and missing functional acknowledgments.
• Saved custom views: Create reusable views with filters and columns for frequently monitored partners and workflows.

The lifecycle view supports X12 and EDIFACT documents, and users can switch back to the transaction view at any time. Lifecycle data for supported documents has been captured since early September, making recent transaction history available without manual reconstruction.

Track EDI documents using the lifecycle view

Automatically detect EDI delimiters

Inbound EDI processing can now detect element, component, and segment separators from incoming X12 and EDIFACT documents.

When a trading partner sends a file with non-default delimiters, Celigo reads the values from the document’s envelope and applies them during parsing. This allows a file definition to process partner-supported documents with different separators, reducing manual configuration changes and parsing errors.

Detection applies to supported version 2 file definitions. Saved file definitions remain unchanged, and outbound EDI generation continues using the configured delimiters.

API management

Define and enforce stronger API contracts

API Builder is adding more detailed request and response schema rules, making it easier to define precise API contracts and validate data without custom scripts.

You can:

• Define required fields, allowed values, patterns, string lengths, and numeric ranges through the visual editor or JSON schema.
• Publish these rules in the generated OpenAPI specification for developers and AI agents consuming your APIs.
• Enable request and response validation for individual APIs.
• Reject invalid requests with HTTP 400 before API execution.
• Identify invalid responses with HTTP 500, with field-level details in the request log and test runs.

Return errors immediately when API connections are offline

API Builder can now return an immediate HTTP 500 response when a connection referenced by an API is offline, rather than waiting up to 2 minutes for a timeout. This allows calling applications to retry or follow their failover logic sooner.

The setting is enabled by default for new APIs and disabled for existing APIs. When enabled, an offline connection causes the API request to fail even if that connection is used only in a branch the current request would not execute.

Connectivity

Improve on-premise integration reliability with agent groups

On-premises integrations no longer need to rely on a single agent running on a single host.

With agent groups, you can configure up to three agents and point a connection at the group instead of an individual agent. If an agent becomes unavailable, Celigo routes new work to another healthy group member.

• Automatic failover: Route new work to the next healthy group member in the configured order.
• Connection-level configuration: Associate existing connections with an agent group without rebuilding the flows that use them.
• Health visibility: Review group membership, agent status, and whether the group can accept work.

Agent groups provide failover rather than load balancing. High availability requires at least two agents, preferably on separate hosts.

Learn more about on-premise agents

Prepare Shopify integrations for expiring access tokens

Shopify is requiring public applications to use expiring offline access tokens beginning January 1, 2027.

Celigo is migrating supported Shopify public-app connections from permanent tokens to access tokens that expire after 60 minutes and are automatically renewed in the background. For most customers, Celigo manages the migration without requiring a reinstall, reauthorization, or changes to existing flows.

FAQs: Shopify expiring offline access token migration

Use customer-owned Shopify OAuth applications

Self-service Shopify OAuth 2.0 connections now support applications created and owned by the customer. This gives teams control over application credentials and OAuth scopes, while providing a separate token chain for each customer-owned application.

Custom applications can also support certain scenarios where the same Shopify store must connect through multiple integrator.io accounts. The capability applies to supported self-service HTTP and GraphQL Shopify connections. Integration app connections continue using Celigo’s public application.

Set up an OAuth 2.0 connection to Shopify

Additional connectivity improvements

NetSuite OAuth 2.0 authentication

Authenticate supported NetSuite connections using OAuth 2.0 Client Credentials, with certificates created and renewed by Celigo, ahead of NetSuite’s 2027.1 release, which blocks new token-based integrations. Existing token-based authentication connections continue to work. OAuth 2.0 does not yet support NetSuite file and blob operations, so flows that need them stay on token-based authentication.

Salesforce execution logs in APIs and tools

Review Salesforce requests, responses, and errors for supported operations within API Builder and Tool Builder. Execution detail follows the applicable logging level. Credentials, OAuth tokens, and session IDs are never logged, and other secrets are masked.

MySQL real-time change data capture

Stream inserts, updates, and deletes from supported MySQL databases using the binary log instead of scheduled polling. Events can include the operation and the row’s previous and updated values. The listener can resume from its last recorded position after a restart, provided the relevant binary logs remain available.

Learn more

Amazon Redshift optimized bulk loading

Load large datasets into Redshift using its native COPY command instead of row-by-row SQL inserts. Celigo manages encrypted, temporary staging, and optional primary keys support upserts. Existing imports retain their current configuration.

Learn more

Azure Synapse Analytics optimized bulk loading

Load large datasets into Azure Synapse dedicated SQL pools through native COPY INTO operations. Records are staged in the customer’s Azure Data Lake Storage Gen2 container. The capability requires service principal authentication, supported storage permissions, and Mapper 2.0.

Google Cloud Storage OAuth 2.0 authentication

Connect to Google Cloud Storage through OAuth 2.0 sign-in as an alternative to downloaded service account credentials. Existing service account connections remain supported.

Learn more

Flexible JSON and XML file templates

File imports now evaluate JSON and XML Handlebars templates as written, without requiring a header, footer, or each block. Existing templates keep working.

Learn more

Safer default file names

New file imports created in the UI default to file names using epoch-based timestamps rather than timestamps containing colons. This improves compatibility with file providers such as Celigo Storage and Azure Data Lake. Existing imports retain their configured file names.

Integration apps and templates

New: Ora-guided solutions in the Marketplace

Discover Ora-guided solutions in the Celigo Marketplace — an AI-assisted experience where Celigo Ora helps you set up and customize integrations through natural-language conversation.

Get ahead of Amazon stockouts and settlement variances

Celigo AI assist introduces two workflows within the Amazon–NetSuite integration app to help teams identify inventory risks and investigate settlement discrepancies.

Inventory forecasting and exception analysis

• Forecast when FBA and merchant-fulfilled SKUs may run out.
• Identify inventory risks associated with upcoming sales events, including Prime Day and Black Friday.
• Flag FBA exceptions, such as stockouts or unfulfillable inventory, and provide explanations and recommended checks.

Settlement variance investigation

• Identify settlements that do not reconcile.
• Determine likely causes of discrepancies and explain relevant Amazon fees.
• Recommend corrective actions in the order they should be addressed.
• Highlight recurring issues that may require additional attention.

Results appear in the Celigo AI assist dashboard in NetSuite, with Inventory forecast and Settlements views. AI assist does not automatically modify Amazon listings, inventory quantities, or underlying financial transactions.

Use Celigo AI assist for Amazon–NetSuite inventory and settlements

Connect Shopify with Microsoft Dynamics 365 Business Central

A new prebuilt template gives Business Central customers a starting point for integrating Shopify direct-to-consumer and B2B commerce workflows.

The template includes:

• 43 prebuilt flows across 11 modules, covering products, orders, inventory, fulfillment, billing, refunds, returns, cancellations, and Shopify Payments payouts.
• B2B support for Shopify Plus, including companies, contacts, and catalogs.
• Configurable commerce settings, including taxes, shipping, payment methods, discounts, multiple locations, and multiple entities.

The template is available in the Integration Marketplace and supports Business Central online deployments. B2B functionality requires Shopify Plus.

Additional developer enhancements

Catch configuration problems before deploying with the Celigo CLI

The Celigo CLI now performs additional validation before pushing or promoting integration resources. Checks help identify broken script references, invalid syntax, missing lookups, invalid topic bindings, and other configuration issues before changes are applied.

Builders can run validation locally or incorporate it into CI workflows, helping developers and AI-assisted development tools detect problems earlier. The CLI also adds support for working with Topics during promotion and introduces additional resource-management commands.

Review the CLI release notes

Explore the complete release

The October and November release cycle includes additional connector updates, platform improvements, technical enhancements, and fixes.

For the complete list of changes and configuration details, visit the Celigo changelog.