OpenAI
Enter a valid instance URL (https:// is required unless it is localhost).
cinatra install cinatra-ai/openai-connector
OpenAI
Connect OpenAI so every agent in your workspace can run on GPT models. Holds your API key, default model, and service tier behind a single configured connection, and ships an opt-in sandboxed shell-tool runtime that lets selected agents execute commands inside a locked-down container during a run.
Works with
- Cinatra AI platform
- OpenAI Responses API (GPT models, e.g.
gpt-5.5) via your own API key
Capabilities
- Run agents on the OpenAI Responses API with your own API key, project, and organization scope
- Pick the default OpenAI model and service tier used across the workspace
- Browse and filter the OpenAI models available to your key
- Validate and store credentials, with optional Nango-backed credential storage
- Write redactable request/response logs for debugging
- Give selected agents a sandboxed Docker shell tool with configurable CPU, memory, network, and command policies
- Mount catalogue skills into the sandboxed shell so agents can pick them up at run time
- Full configuration, usage, and architecture docs in docs/USAGE.md
- Converted the connector to the schema-config surface: removed the custom React configuration pages; configuration now renders from a declared config schema and configures fully at runtime with no image rebuild.
- Declared the connector's supported Cinatra SDK ABI range so the in-instance compatibility badge reads Compatible.
More from Cinatra
Plug Apify's actor catalogue into every Cinatra agent. Once connected, any agent that needs to scrape, crawl, or extract data can reach for an…
Enter a valid instance URL (https:// is required unless it is localhost).
cinatra install cinatra-ai/apify-connector
Run the platform's policy scanners against an agent definition and report what they find. The agent vets a definition for hardcoded secrets, untrusted URLs,…
Enter a valid instance URL (https:// is required unless it is localhost).
cinatra install cinatra-ai/lint-policy-agent
Let people choose which workspace artifacts feed into another agent's context before it runs. When a larger agent needs supporting material — an ICP,…
Enter a valid instance URL (https:// is required unless it is localhost).
cinatra install cinatra-ai/context-selection-agent
Connect Google's Gemini API so Cinatra can use Gemini models for transcription, image generation, and general-purpose generation. Once connected, agents and workflows pinned to…
Enter a valid instance URL (https:// is required unless it is localhost).
cinatra install cinatra-ai/gemini-connector
A slide-format presentation exported to PDF — pitch decks, sales decks, conference talks, board decks, training presentations, and customer-facing decks. Slide Decks get a…
Enter a valid instance URL (https:// is required unless it is localhost).
cinatra install cinatra-ai/slide-deck-artifact
The system audio player for the Cinatra artifact library. It renders any audio file — MP3, M4A, Ogg, WAV, WebM, FLAC, or AAC —…
Enter a valid instance URL (https:// is required unless it is localhost).
cinatra install cinatra-ai/audio-artifact
Reviews
There are no reviews yet.