gamma-sdk-patterns

Output

Produce a client-validation receipt with contract/fixture version, opaque operation ID, schema result, idempotency outcome, owner, and redacted failure reference. Do not log presentation content or tokens.

Examples

Use a scoped staging integration and fictional document to validate field allowlists and duplicate suppression. Reject an unknown destination before the provider call and retain only aggregate results.

Overview

Gamma has no published SDK — all interaction is via REST at https://public-api.gamma.app/v1.0/. This skill provides production-grade patterns for typed clients, generation helpers, polling, template workflows, and error handling.

Prerequisites

• Completed gamma-install-auth setup
• TypeScript project with fetch (Node.js 18+)
• Understanding of the generate-poll-retrieve workflow

Instructions

Step 1: Typed Client Singleton

// lib/gamma.ts
const GAMMA_BASE = "https://public-api.gamma.app/v1.0";

interface GammaConfig {
  apiKey: string;
  baseUrl?: string;
  timeoutMs?: number;
}

// Types based on actual API responses
interface GenerateRequest {
  content: string;
  outputFormat?: "presentation" | "document" | "webpage" | "social_post";
  themeId?: string;
  exportAs?: "pdf" | "pptx" | "png";
  textMode?: "generate" | "condense" | "preserve";
  textAmount?: "brief" | "medium" | "detailed" | "extensive";
  imageOptions?: { style?: string };
  sharingOptions?: {
    workspaceAccess?: "noAccess" | "view" | "comment" | "edit" | "fullAccess";
    externalAccess?: "noAccess" | "view" | "comment" | "edit" | "fullAccess";
  };
  folderIds?: string[];
}

interface GenerateResult {
  generationId: string;
  status: "in_progress" | "completed" | "failed";
  gammaUrl?: string;
  exportUrl?: string;
  creditsUsed?: number;
}

let instance: ReturnType<typeof createGammaClient> | null = null;

export function getGamma() {
  if (!instance) {
    instance = createGammaClient({
      apiKey: process.env.GAMMA_API_KEY!,
    });
  }
  return instance;
}

export function createGammaClient(config: GammaConfig) {
  const base = config.baseUrl ?? GAMMA_BASE;
  const headers: Record<string, string> = {
    "X-API-KEY": config.apiKey,
    "Content-Type": "application/json",
  };

  async function request<T>(method: string, path: string, body?: unknown): Promise<T> {
    const controller = new AbortController();
    const timeout = setTimeout(() => controller.abort(), config.timeoutMs ?? 30000);
    try {
      const res = await fetch(`${base}${path}`, {
        method,
        headers,
        body: body ? JSON.stringify(body) : undefined,
        signal: controller.signal,
      });
      if (!res.ok) {
        const text = await res.text();
        throw new GammaApiError(res.status, text, path);
      }
      return res.json() as T;
    } finally {
      clearTimeout(timeout);
    }
  }

  return {
    generate: (body: GenerateRequest) =>
      request<{ generationId: string }>("POST", "/generations", body),
    generateFromTemplate: (body: TemplateRequest) =>
      request<{ generationId: string }>("POST", "/generations/from-template", body),
    poll: (id: string) =>
      request<GenerateResult>("GET", `/generations/${id}`),
    getFileUrls: (id: string) =>
      request<{ exportUrl: string }>("GET", `/generations/${id}/files`),
    listThemes: () => request<Theme[]>("GET", "/themes"),
    listFolders: () => request<Folder[]>("GET", "/folders"),
  };
}

Step 2: Custom Error Class

// lib/errors.ts
export class GammaApiError extends Error {
  constructor(
    public status: number,
    public body: string,
    public path: string
  ) {
    super(`Gamma API ${status} on ${path}: ${body}`);
    this.name = "GammaApiError";
  }

  get isRateLimit() { return this.status === 429; }
  get isAuth() { return this.status === 401 || this.status === 403; }
  get isServerError() { return this.status >= 500; }
}

Step 3: Poll-Until-Done Helper

// lib/poll.ts
export async function pollUntilDone(
  gamma: ReturnType<typeof createGammaClient>,
  generationId: string,
  opts = { intervalMs: 5000, timeoutMs: 180000 }
): Promise<GenerateResult> {
  const deadline = Date.now() + opts.timeoutMs;

  while (Date.now() < deadline) {
    const result = await gamma.poll(generationId);

    if (result.status === "completed") return result;
    if (result.status === "failed") {
      throw new Error(`Generation ${generationId} failed`);
    }

    await new Promise((r) => setTimeout(r, opts.intervalMs));
  }

  throw new Error(`Poll timeout for ${generationId} after ${opts.timeoutMs}ms`);
}

Step 4: Generate-and-Wait Convenience

// lib/generate.ts
export async function generateAndWait(
  gamma: ReturnType<typeof createGammaClient>,
  request: GenerateRequest
): Promise<GenerateResult> {
  const { generationId } = await gamma.generate(request);
  console.log(`Generation started: ${generationId}`);
  return pollUntilDone(gamma, generationId);
}

// Usage
const gamma = getGamma();
const result = await generateAndWait(gamma, {
  content: "Quarterly business review for Q1 2026",
  outputFormat: "presentation",
  themeId: "theme_abc123",
  exportAs: "pptx",
  textAmount: "medium",
  imageOptions: { style: "photorealistic corporate" },
});
console.log(`View: ${result.gammaUrl}`);
console.log(`Download: ${result.exportUrl}`);

Step 5: Template-Based Generation

// lib/templates.ts
// Uses POST /v1.0/generations/from-template
// The template gamma must contain exactly one page

interface TemplateRequest {
  gammaId: string;     // Template gamma ID (one-page template)
  prompt: string;      // Content + instructions for the template
  themeId?: string;
  exportAs?: "pdf" | "pptx" | "png";
  imageOptions?: { style?: string };
  sharingOptions?: object;
  folderIds?: string[];
}

export async function generateFromTemplate(
  gamma: ReturnType<typeof createGammaClient>,
  templateId: string,
  prompt: string,
  options: Partial<TemplateRequest> = {}
): Promise<GenerateResult> {
  const { generationId } = await gamma.generateFromTemplate({
    gammaId: templateId,
    prompt,
    ...options,
  });
  return pollUntilDone(gamma, generationId);
}

Step 6: Retry with Backoff

// lib/retry.ts
export async function withRetry<T>(
  fn: () => Promise<T>,
  maxRetries = 3,
  baseDelayMs = 1000
): Promise<T> {
  for (let attempt = 0; attempt <= maxRetries; attempt++) {
    try {
      return await fn();
    } catch (err) {
      if (attempt === maxRetries) throw err;
      if (err instanceof GammaApiError && !err.isRateLimit && !err.isServerError) {
        throw err; // Don't retry auth errors or 4xx
      }
      const delay = baseDelayMs * Math.pow(2, attempt);
      console.warn(`Retry ${attempt + 1}/${maxRetries} in ${delay}ms`);
      await new Promise((r) => setTimeout(r, delay));
    }
  }
  throw new Error("Unreachable");
}

// Usage
const result = await withRetry(() =>
  generateAndWait(gamma, { content: "My deck", outputFormat: "presentation" })
);

API Endpoints Reference

| Method | Endpoint | Purpose | |--------|----------|---------| | POST | /v1.0/generations | Generate from text content | | POST | /v1.0/generations/from-template | Generate from a template gamma | | GET | /v1.0/generations/{id} | Poll generation status | | GET | /v1.0/generations/{id}/files | Get export file URLs | | GET | /v1.0/themes | List workspace themes | | GET | /v1.0/folders | List workspace folders |

Error Handling

| Pattern | Use Case | |---------|----------| | GammaApiError class | Typed error handling with isRateLimit, isAuth, isServerError | | withRetry() | Auto-retry on 429/5xx with exponential backoff | | pollUntilDone() | Timeout-aware polling with configurable interval | | Singleton getGamma() | Consistent config across modules |

Resources

• Gamma API Reference
• Generate API Parameters
• Create from Template

Next Steps

Proceed to gamma-core-workflow-a for content generation workflows.