Pipedream OpenAI → n8n

7 Pipedream OpenAI actions and 4 event sources, and the n8n nodes they become. Converted nodes keep your configured values; expressions like {{steps.x.$return_value}} are rewritten to n8n expressions. Pipedream app slug: openai.

Actions

Pipedream action	n8n	Fit	Notes
openai-chat	Open Ai text → response	Check output	API change: PD calls POST /v1/chat/completions; n8n OpenAI V2 'Message a Model' (operation value 'response') calls POST /v1/responses. Equivalent for plain prompting, but: no n/stop/presence/frequency penalties; audio input unsupported; PD function tools…
openai-chat-with-responses-api	Open Ai text → response	Check output	Same endpoint (POST /v1/responses). otherTools (raw tool JSON e.g. remote MCP servers {type:'mcp', server_url,...}) has no parameter in the n8n node; attach an MCP Client Tool / other tool sub-node to the node's ai_tool input instead (n8n executes tools…
openai-create-image	Open Ai image → generate	Check output	n8n V2 >= 2.2 uses modelId resourceLocator (V2 < 2.2 used options param `model`), and options collection `imageGenerateOptionsRLC` (same name 'options'). n8n `dalleQuality` is renamed to body.quality at runtime; options are spread into the request body. PD…
openai-create-transcription	Open Ai audio → transcribe	Check output	n8n transcribe hardcodes model 'whisper-1' (no model, prompt, response_format, include or timestamp_granularities params), so gpt-4o-transcribe / gpt-4o-mini-transcribe selections are lost. If those matter use HTTP Request POST…
openai-create-embeddings	Http Request	Check output	HTTP fallback: the n8n OpenAI V2 node has no embeddings operation (embeddingsOpenAi is an AI sub-node only). PD validates each input element <= 8192 chars client-side; PD declares `user` but never sends it, so it is dropped. If the openAiApi credential uses…
openai-analyze-image-content	Open Ai image → analyze	Check output	Same endpoint (POST /v1/responses with input_text + input_image). n8n sends max_output_tokens = options.maxTokens \|\| 300 — PD sends none, so 4096 is set here to avoid truncation (no PD equivalent; adjust if needed). PD uploads /tmp files to /files (purpose…
openai-send-prompt	Http Request	Check output	HTTP fallback: legacy /v1/completions (only gpt-3.5-turbo-instruct, davinci-002, babbage-002) has no operation in the n8n OpenAI node. PD declares `suffix` but _getCommonArgs never sends it. `alert` prop is display-only.

Triggers (event sources)

Pipedream trigger	n8n	Fit	Notes
openai-new-batch-completed	Schedule Trigger + polling	Partial	No OpenAI trigger node in n8n. Recommend Schedule Trigger (interval = PD timer) -> HTTP Request GET https://api.openai.com/v1/batches (nodeCredentialType openAiApi, paginate via `after`) -> Split Out `data` -> filter completed_at > last run (workflow static…
openai-new-file-created	Schedule Trigger + polling	Partial	No n8n trigger. Schedule Trigger -> OpenAI node resource 'file' operation 'list' (or HTTP GET https://api.openai.com/v1/files?purpose=<purpose>) -> dedupe on id. includeLink/File Stash has no equivalent.
openai-new-fine-tuning-job-created	Schedule Trigger + polling	Partial	No n8n trigger. Schedule Trigger -> HTTP Request GET https://api.openai.com/v1/fine_tuning/jobs (openAiApi) -> dedupe on id.
openai-new-run-state-changed	Schedule Trigger + polling	Partial	No n8n trigger. Schedule Trigger -> HTTP Request GET https://api.openai.com/v1/threads/<threadId>/runs with header OpenAI-Beta: assistants=v2 (openAiApi) -> Code node comparing statuses with workflow static data. Assistants API is deprecated by OpenAI.

Credentials

Create a OpenAI credential in n8n (type openAiApi). Converted nodes reference it by name, so n8n links it on import when the names match. Pipedream connected accounts can't be exported and are deleted by April 30, 2027.

What "Fit" means

Direct: same operation, same inputs. Check output: the n8n node returns a different JSON shape, so later steps that read its result need a look. Partial: some options have no n8n equivalent. Mappings verified on 2026-10-13 against the Pipedream registry and n8n source.

Questions

Does n8n have a OpenAI node?+

Yes. 5 of the 7 Pipedream OpenAI actions listed here map to a native n8n node; 2 map to n8n's HTTP Request node with the OpenAI credential.

Will my Pipedream OpenAI connection move to n8n?+

No. Pipedream exports only an account ID, never the token, and deletes connected accounts by April 30, 2027. Create a OpenAI credential in n8n (type openAiApi) and select it on the converted nodes.

Are Pipedream OpenAI triggers supported in n8n?+

n8n has no trigger node for these OpenAI events. Converted workflows use a Schedule Trigger that polls the API, or a Webhook node registered in OpenAI.

What happens to code steps that call the OpenAI API?+

Code steps keep working through a compatibility layer, and tokens read via this.<app>.$auth move to a secrets block in the Code node. Steps that use an npm SDK are rewritten to plain HTTP calls so they run on n8n Cloud.

Other apps

Airtable Asana Calendly ClickUp Discord Dropbox Email (SMTP)GitHub Gmail Google Calendar Google Drive Google Gemini All apps →