Ngram
Ngram creates AI-generated videos and images, manages rendering jobs, and delivers completion events through webhooks.
Verdict
Common use cases
- Prototype hero images for landing pages
- Generate social media video from campaign copy
- Mock product screenshots before design handoff
- Check credit spend before bulk render jobs
- Cancel accidental video renders mid-flight
Integration
- Vendor
- Ngram
- Category
- other
- Auth
- API_KEY
- Tools
- 14
- Composio slug
ngram
Tools
- Cancel Video
Irreversibly request cancellation of an in-progress Ngram video render. Confirm the target video before calling. Ngram states that cancellation does not refund work already completed.
- Create Webhook Subscription
Subscribe a public HTTPS endpoint to all account-wide Ngram video completion or failure events of one type. The response includes a signing secret that is shown only at creation and must be retained to verify webhook signatures.
- Delete Webhook Subscriptiondestructive
Delete an Ngram webhook subscription by ID. The operation is idempotent: a missing subscription returns deleted=false rather than an error.
- Generate Image
Spend Ngram credits to start one image generation job from a prompt, optionally using one to four public HTTPS reference images for transformation. Inspect NGRAM_GET_CREDITS and NGRAM_GET_RENDER_OPTIONS first, then deliberately set confirm_
- Generate Video
Spend Ngram credits to start one asynchronous video render from text or a public URL. Inspect NGRAM_GET_CREDITS and NGRAM_GET_RENDER_OPTIONS first, then deliberately set confirm_credit_spend=true and supply a stable idempotency_key. Exact p
- Get Account
Return the authenticated Ngram account identity, including its id, email, and name.
- Get Credits
Return the connected account's plan and available, reserved, and effective credits. Use effective credits as the spendable balance before generation.
- Get Image
Return an image job's current status and generated output by its UUID.
- Get Render Options
Return Ngram's live catalog of supported video and image capabilities, voices, styles, formats, durations, models, templates, and other generation options. Use this before generation instead of assuming catalog values.
- Get Video
Return a video's render status, progress, output URL, or failure information by video ID.
- List Brand Kits
Return the brand kits available to the connected account for use in generation requests.
- List Images
Return up to 100 account image jobs matching an optional status and created-at upper bound. Ngram does not expose a continuation token.
- List Videos
Return up to 100 account videos matching an optional render status and created-at lower bound. This is one bounded response; Ngram provides no continuation token.
- List Webhook Subscriptions
Return the account's video completion and failure webhook subscriptions. Signing secrets are not returned by this list endpoint.
Setup
Setup guide
- 11. Open your Switchy workspace settings and navigate to the Integrations tab. 2. Click Add MCP Integration and select Ngram from the directory. 3. Log in to your Ngram account dashboard and generate a new API key under Account → API Keys. 4. Paste the key into Switchy's prompt and click Connect. 5. Switchy will confirm the connection by fetching your account identity and current credit balance. 6. Inside any Project, type '@Ngram what's my credit balance?' to verify the integration responds. 7. To queue a render, write '@Ngram generate an image: a cat wearing sunglasses' and the MCP will return a job UUID you can poll for completion. 8. If your team needs webhook notifications for finished renders, use the Create Webhook Subscription tool and point it at a public HTTPS endpoint you control; Ngram will POST events there with a signing secret shown only once.
What teammates see: by default, memories from Ngram are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Projects have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Check Credit Balance
@Ngram show my current credit balance and tell me how many are reserved versus available to spend right now.
Generate Hero Image
@Ngram generate an image: a modern SaaS dashboard with clean typography and a gradient background, 16:9 aspect ratio.
Render Marketing Video
@Ngram generate a 10-second video: a product unboxing on a white table with soft lighting. Confirm credit spend and return the job ID.
Poll Render Status
@Ngram get the status of image job <UUID> and show me the generated output if it's complete.
Cancel In-Progress Video
@Ngram cancel video job <UUID> immediately and confirm the cancellation went through.
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@ngram check my credit balance and tell me how many video renders I can afford at standard quality
Your Ngram account shows 450 effective credits available. Standard-quality video renders typically cost 50–100 credits per job depending on duration and complexity. Based on current pricing, you can afford approximately 4–9 standard renders before needing to top up. Reserved credits (120) are held for in-progress jobs and will return to your balance when those complete.
This example demonstrates credit-balance inspection before spending. Ngram charges per render and does not refund cancelled jobs for work already completed, so checking available credits first prevents mid-project budget surprises. Requires valid API_KEY authentication.
@ngram generate a 10-second video of a cat riding a skateboard through a neon-lit cyberpunk alley, confirm I want to spend credits
Video render started. Job ID: vid_8k3m9p2x. Ngram is processing your prompt asynchronously; typical completion time is 3–8 minutes for a 10-second clip. You've been charged approximately 75 credits. Poll the job status with the ID above, or set up a webhook subscription to receive a notification when rendering finishes or fails.
This write-action example shows the two-step safety pattern: the MCP requires confirm_credit_spend=true to prevent accidental billing. Video generation is async—immediate response is a job ID, not the final asset. Users should inspect render options and credit cost before committing to large batches.
@ngram list all my in-progress video jobs and cancel any that have been running longer than 15 minutes
Found 2 in-progress renders. Job vid_7a4n1k9s (started 18 minutes ago, prompt: 'drone flyover of mountain range') has been cancelled; Ngram will not refund the 60 credits already consumed. Job vid_2p8m5x1r (started 4 minutes ago) remains active. No other jobs meet the 15-minute threshold.
This synthesis example pairs Ngram's job-listing and cancellation tools with time-based logic. Cancellation is irreversible and does not recover spent credits, so the AI's reasoning layer helps avoid mistaken cancellations. Useful for cleaning up stalled renders or managing account activity at scale.
Use-case deep-dives
When Ngram fits a weekly batch content workflow
A three-person marketing team publishes 15 social posts per week across LinkedIn and Instagram. They write captions in Notion on Monday, then need matching visuals by Wednesday. Ngram's image and video generation tools let them batch-create assets from prompts without leaving their workspace, and the webhook subscription means they can queue 15 jobs Monday night and get completion notifications by morning. The credit system works here because spend is predictable and monthly. If your team publishes daily or needs real-time generation during client calls, the async model adds friction. For weekly batches under 50 assets, Ngram keeps the creative pipeline moving without a separate design tool.
Ngram scales personalized video at small SaaS scale
A six-person customer success team at a B2B SaaS company sends personalized onboarding videos to 40 new accounts per month. Each video includes the customer's logo and a custom walkthrough script. Ngram's reference image transformation and text-to-video tools let the team template the structure once, then generate variants by passing account-specific prompts and logo URLs. The webhook subscription notifies Slack when each render completes, so the CSM can review and send same-day. Credit spend stays under $200/month at this volume. If you onboard more than 100 accounts monthly, the per-video cost and manual review step become bottlenecks. Below that threshold, Ngram delivers personalization that would otherwise require a video editor on retainer.
When Ngram replaces stock photo hunting for decks
A four-person agency updates client pitch decks every quarter, replacing stale hero images and section breaks with on-brand visuals. Ngram's image generation from prompts eliminates the two-hour Unsplash hunt and the licensing ambiguity. The team writes prompts that match each deck's tone, generates three options per slide, and picks winners in under 30 minutes. The get-credits tool shows remaining balance before they commit to a batch, preventing mid-sprint budget surprises. If your decks need photo-realistic product shots or exact brand color matching, Ngram's output won't pass legal review. For concept visuals and abstract backgrounds in early-stage pitches, it closes the deck faster than any stock library.
Frequently asked
What does the Ngram MCP do in Switchy?
It lets your team generate AI images and videos directly from chat by calling Ngram's API. You can start renders from text prompts or reference URLs, check credit balances, cancel in-progress jobs, and subscribe webhooks to track completion events. All generation costs draw from your Ngram account's credit pool.
Do I need an Ngram API key to connect this MCP?
Yes. The MCP uses API key authentication, so whoever connects it must paste a valid Ngram API key from their account settings. That key ties all generation requests and credit spending to a single Ngram account, so choose the team member who owns the billing relationship.
Can the Ngram MCP refund credits if I cancel a video mid-render?
No. Ngram's own API states that cancellation does not refund work already completed. The MCP's Cancel Video tool will stop the job, but any credits consumed up to that point are gone. Check your effective credit balance with Get Credits before starting expensive renders.
Should I use this MCP or call Ngram's API directly in my app?
Use the MCP if your team wants to prototype renders in Switchy chat without writing code. Use the API directly if you need programmatic control, batch workflows, or tighter error handling. The MCP is a thin wrapper—it won't add features Ngram's API doesn't expose.
Who on my team should connect the Ngram MCP?
Whoever holds the Ngram account with credits you're willing to spend. Every image or video generation triggered through Switchy will deduct from that account's balance. If multiple people need access, consider a shared Ngram team account or rotate the API key when the credit owner changes.