Ignition 8.3.9: What the New API Updates Mean for AI Agents
Ignition 8.3.9 adds EAM batch APIs and Edge Historian routes. Here is what to update in your AI agent's tools, Jython instructions, and project checks.
Key takeaways
- The new API routes cover EAM batch operations and Edge Historian statistics.
- Use the target Gateway's OpenAPI description to update agent tool definitions.
- Jython 2.7.4 arrived in 8.3.8. Ignition 8.3.9 still requires Python 2.7-compatible scripts.
- Some historian parameters are still accepted but ignored. Check the returned data against a known result.
- Revisit generated Perspective views, relative tag references, and migration behavior after upgrading.
Related product pages
What changed in Ignition 8.3.9?
Ignition 8.3.9 adds API routes for EAM batch operations and Edge Historian statistics. If you use an AI agent to help develop or manage Ignition projects, those additions are worth reviewing alongside the release's scripting changes. Inductive Automation's release notes describe the changes.
An AI agent can use APIs through its connected tools, with the same access requirements as any other client. The useful question is what changed in Ignition and how your agent's tools and instructions should change with it.
EAM and Edge Historian get new API routes
In EAM, an agent is a managed Ignition Gateway. An AI assistant could help an engineer prepare a Gateway upgrade plan or collect historian statistics for troubleshooting. For a batch operation, have it record the intended Gateways and report the outcome for each one. Accepting the request is only the first step. EAM documentation explains the controller and agent relationship.
The new API work covers:
- Bulk approval and deletion of EAM agents.
- Batch upgrades of EAM agents.
- Routes that expose Edge Historian statistics.
Refresh the Gateway's OpenAPI description
Start with /openapi and /openapi.json on the target Gateway. Ignition generates the specification from the Gateway and its installed modules. It describes the public operations available on that installation, including their inputs, authentication requirements, and responses. Internal and session-dependent routes are excluded from that description. Ignition's OpenAPI documentation covers the details.
Use that specification when updating the agent's tool definitions. Check the HTTP method, request fields, required permissions, and response format. For an existing integration, compare specifications exported before and after the upgrade, using the same module configuration.
This gives the agent a concrete reference for constructing requests and interpreting results. It also gives the next engineer a record of what changed.
Keep the Jython 2.7 rules
Jython 2.7.4 arrived in Ignition 8.3.8. It remains the scripting baseline in 8.3.9. The language still follows Python 2.7 rules, so generated Ignition scripts should use compatible formatting such as .format(). F-strings and async/await do not belong in those scripts. See the Ignition version history and the Jython project.
Claude or Codex can run outside Ignition while generating code for its embedded runtime. Tell the agent which Ignition version and script scope it is targeting. Then test the script where it will run, whether that is a Perspective event, a Gateway event, or the Designer Script Console.
A historian query can run and still give you the wrong answer
Ignition 8.3.9 removes unsupported historian parameters from autocomplete, IDE hints, and documentation. Existing scripts can still pass them, but Ignition continues to ignore them at runtime. These were already unsupported parameters. The release does not remove previously working behavior. The release notes identify these arguments:
system.historian.queryAggregatedPoints(): fillModes, includeBounds, and excludeObservations.
system.historian.queryRawPoints(): includeBounds and excludeObservations.
system.historian.browse(): snapshotTime.Test the returned data
An agent might add fillModes, get a dataset back, and tell you the query now applies the interpolation you requested. The argument had no effect.
Give it a small known dataset and an expected answer instead. Have it identify the historian provider and compare timestamps, columns, quality, and values. Keep separate expectations for SQL, Core, and Internal Historians. Run those checks on development or staging.
Put the release details in the skill file
A skill file is a place to keep instructions the agent needs on every task. For 8.3.9, that includes the target version, API reference, Jython rules, historian constraints, and expected checks. Put them there so the next task starts with the same engineering requirements.
Start with read-only discovery and a proposed diff. Review the change, use a supported dry-run where available, and apply it on development or staging. Then have the agent read back the saved configuration and test the runtime behavior. Production deployment needs human engineering approval.
What this article establishes
This article is based on official documentation reviewed September 5, 2026. It does not report tests on an 8.3.9 Gateway or a comparison of exported 8.3.8 and 8.3.9 OpenAPI specifications.
Our current public 8.3 packages list Ignition 8.3.8 as their target. This release review does not establish that those packages have been validated on 8.3.9.
Give your AI agent the Ignition references it needs
Take a look at our Ignition 8.3 API enhancements for the additional actions supplied by the toolkit. The API package complements Ignition's built-in OpenAPI and documents the result each action should return.
Our Ignition 8.3 skills cover Jython scripting, Perspective, Vision, expressions, and UDTs. Choose the skills for the work you are doing, check their stated tested version, and validate the workflow on your target Gateway.
Article FAQ
Frequently asked questions
Are there new API endpoints in Ignition 8.3.9?
Yes. The release notes identify EAM endpoints for bulk approval and deletion, an EAM batch upgrade endpoint, and routes exposing Edge Historian statistics. Inspect the target Gateway's OpenAPI description to establish which operations are available on that installation.
Can an AI agent use Ignition's APIs?
An AI agent can use an API through its connected tools when those tools have network access, the required authentication, and permission for the operation. API access depends on the installation and tool integration, just as it does for other clients.
Does Ignition 8.3.9 introduce Jython 2.7.4?
No. Jython 2.7.4 arrived in Ignition 8.3.8 and remains the scripting baseline in 8.3.9. It follows Python 2.7 syntax, so f-strings and async/await are not valid choices for generated Ignition scripts.
Where should an AI agent get the Ignition OpenAPI specification?
Use /openapi and /openapi.json on the target Gateway. Ignition generates the description from that Gateway and its installed modules. Compare exports made before and after an upgrade using the same module configuration.
Sources and notes
Documentation referenced
Lifetime membership
Want the toolkit behind this workflow?
The one-time Ignition AI Toolkit membership includes the Web Dev API runner, skill files, setup docs, versioned downloads, and released Ignition 8.3 resources for lifetime members. Founder pricing is available while the first 50 spots last.
Founder lifetime access - $99