Documents, OCR and presentations APIs and MCP servers for AI agents

Parse PDFs and scans, convert file formats, generate slides and collect e-signatures.

14 documents, ocr and presentations APIs an AI agent can call: auth method and header, official MCP server (10), free tier (8), docs links. Machine-readable JSON included. Agents: fetch catalog/presentations-documents.json instead of parsing this page. Last checked 2026-09-24.

Ratings and usage (★) fetched: 2026-09-24 16:54 UTC – 2026-09-24 17:02 UTC.

Service	What an agent can do	Auth	MCP server	Free tier	Details (with sources)	Notes	Proof (ratings, usage)
Adobe PDF Services API	Programmatically create, convert, combine, compress, OCR, and extract data from PDF documents.	OAuth 2.0 `Authorization: Bearer <access_token> obtained via OAuth Server-to-Server client_credentials grant (client_id/client_secret)`	—	free tier: 500 document transactions/month, no credit card required	Base URL: https://pdf-services.adobe.io Free plan: 500 Document Transactions, per month ↗ Rate limits: Free tier 25 RPM, Enterprise 100 RPM; max file size 100MB ↗ Async jobs: job creation returns a location header; poll it until status done/failed ↗ Example call: OCR an uploaded PDF: POST https://pdf-services.adobe.io/operation/ocr ↗	As of January 2025, JWT/service-account auth is deprecated; all new integrations must use OAuth Server-to-Server (2-legged client-credentials) via Adobe Developer Console.	—
AWS Textract	Detect and extract text, forms, tables, and data from scanned documents and images via API calls.	cloud IAM `AWS SigV4-signed requests using IAM access key/secret or role credentials`	official	Trial: new AWS customers get 3 months free, per month 1,000 pages DetectDocumentText, 100 pages AnalyzeDocument forms/tables/layout/queries, 100 AnalyzeExpense/AnalyzeID, 2,000 pages AnalyzeLending; card required	Base URL: https://textract.us-east-1.amazonaws.com Free plan: per month for 3 months: DetectDocumentText 1,000 pages; AnalyzeDocument forms/tables/layout 100 pages; AnalyzeExpense/AnalyzeID 100 pages; AnalyzeLending 2,000 pages, card required ↗ Rate limits: us-east-1 defaults: DetectDocumentText 25 TPS, AnalyzeDocument 10 TPS, AnalyzeExpense/AnalyzeID 5 TPS; up to 600 concurrent async jobs ↗ Async jobs: Start* operations (e.g. StartDocumentAnalysis) return a JobId; poll Get* operations; sync APIs also available ↗ Example call: detect text in a document image: POST https://textract.{region}.amazonaws.com/ ↗ Unit price: DetectDocumentText $0.0015 per page (first 1M pages/month); AnalyzeDocument tables $0.015/page, forms $0.05/page (US West Oregon) ↗ MCP tools: 8: aws___run_script, aws___get_presigned_url, aws___get_tasks, aws___search_documentation, aws___read_documentation, aws___retrieve_skill … ↗	Requires an AWS account and IAM credentials; region-specific; free tier applies only for the first 3 months after the first Textract request. No dedicated Textract MCP; the official AWS MCP Server can call Textract via aws___run_script.	GitHub 0★ ag2-mcp-servers/amazon-textract (2026-09-24 16:54 UTC)
Azure AI Document Intelligence	Analyze forms, invoices, receipts, and scanned documents to extract structured JSON via REST API.	cloud IAM `Ocp-Apim-Subscription-Key: <key> header (Azure AD/Entra ID Bearer token also supported)`	—	Free tier: F0 pricing tier, 500 pages/month; analyzes only the first 2 pages per document, 4 MB max file, 1 TPS	Free plan: 500 pages (F0); only the first 2 pages of each document analyzed, 4 MB max, per month ↗ Rate limits: F0: 1 TPS analyze, 1 TPS get; S0: 15 TPS analyze, 50 TPS get (adjustable) ↗ Async jobs: POST ...:analyze returns 202 with Operation-Location header; poll GET until status succeeded/failed ↗ Example call: analyze a document with a prebuilt model (prebuilt-layout, prebuilt-invoice, prebuilt-receipt ...): POST {endpoint}/documentintelligence/documentModels/{modelId}:analyze?api-version=2024-11-30 ↗ Unit price: Pay-as-you-go (East US, S0): prebuilt models $10 per 1,000 pages; Read $1.50 per 1,000 pages; custom extraction $30 per 1,000 pages ↗	Auth via Azure subscription key or Entra ID (Azure AD) OAuth2; current GA version is 4.0; legacy v3.0 API scheduled to retire March 30, 2029; no official Microsoft MCP server specific to Document Intelligence, only community implementations and the general-purpose Azure MCP Server.	—
CloudConvert API	Convert, compress, merge, watermark, and OCR files across 200+ formats via REST API or MCP.	API key / OAuth `Authorization: Bearer <API_KEY> (OAuth2 authorization-code/implicit flow also supported for user-facing apps)`	official	Free: 10 conversion credits/day including all API features, no credit card required; paid credit packages beyond that; free Sandbox API for unlimited test jobs without consuming credits	Base URL: https://api.cloudconvert.com/v2 Free plan: 10 conversion credits/day, all API features included, no card ↗ Rate limits: Dynamic rate limiting on task/job creation; 429 with X-RateLimit-* and Retry-After headers ↗ Async jobs: jobs/tasks run asynchronously; poll job status or receive webhooks ↗ OAuth scopes: user.read, task.read, task.write ↗ Example call: convert a file (import URL, convert, export URL job): POST https://api.cloudconvert.com/v2/jobs ↗ MCP tools: 2: createJob, getOperations ↗	Official MCP server at https://mcp.cloudconvert.com uses Streamable HTTP transport and authenticates via OAuth2 against the user's CloudConvert account (no API key needed for MCP); no public OpenAPI/Swagger file is served by CloudConvert itself.	SourceForge 5.0★/1 review (2026-09-24 16:56 UTC)
DocuSign eSignature API	Send, sign, track, and manage electronic signature envelopes and agreement workflows via REST API.	OAuth 2.0 `Authorization: Bearer <access_token> via OAuth2 Authorization Code grant or JWT Grant`	official	free developer/demo sandbox account (account-d.docusign.com) with unlimited test envelopes; production use requires a paid DocuSign plan	Free plan: non-expiring developer (demo) account; envelopes not legally valid; production from $50/mo Starter (40 envelopes/mo) ↗ OAuth scopes: signature, impersonation ↗ Example call: create and send an envelope for signature: POST https://demo.docusign.net/restapi/v2.1/accounts/{accountId}/envelopes ↗	Official Docusign MCP Server (Open Beta, OAuth) covers eSignature plus Navigator/Maestro/Workspaces: production https://mcp.docusign.com/mcp, developer accounts https://mcp-d.docusign.com/mcp. JWT grant needs one-time user consent.	GitHub 24★ docusign/OpenAPI-Specifications (2026-09-24 16:54 UTC)
Gamma API	Generate, poll, and export AI-made presentations, documents, or social posts as PDF/PPTX via REST API.	API key / OAuth `X-API-KEY: <key> (OAuth 2.0 Bearer tokens also accepted on the same endpoints)`	official	Paid only for API keys (Pro, Ultra, Teams or Business plan); the Gamma MCP server/connectors work on all plans including Free, with generations consuming credits	Base URL: https://public-api.gamma.app/v1.0 Free plan: none ↗ Rate limits: Burst, rolling and daily limits reported via x-ratelimit-remaining-burst / -remaining / -remaining-daily headers ↗ Async jobs: POST /generations returns generationId; poll GET /generations/{id} ↗ OAuth scopes: generate, gamma:read ↗ MCP tools: 17: generate, generate_multi_page_gamma, generate_from_template, generate_image, get_generation_status, get_image_generation_status … ↗	Async workflow (POST /v1.0/generations then poll GET .../{id}); base URL https://public-api.gamma.app/v1.0; official remote MCP server is OAuth2-secured (RFC 7591 dynamic client registration).	—
Google Document AI	Extract text, tables, and structured data from documents like invoices, forms, and IDs via REST API.	cloud IAM `Authorization: Bearer <OAuth2 access token> (service account) or API key for some endpoints`	—	Free tier: first 1,000 pages/month free for Enterprise Document OCR only (then $1.50/1,000); Form Parser and custom extractor $30/1,000 pages with no free band	Base URL: https://documentai.googleapis.com Free plan: first 1,000 pages free (Enterprise Document OCR only); Form Parser/custom extractor billed from page 1 ($30/1,000), per month ↗ Rate limits: Online process: 120 requests/min per project per processor type (US/EU); 5 concurrent batch requests per project ↗ Data policy: Content not used except to provide the service; online requests not persisted; batch data deleted after processing (1-day TTL failsafe) ↗ Async jobs: :process is synchronous; :batchProcess returns a long-running operation to poll ↗ Example call: process a document (OCR / extraction): POST https://{LOCATION}-documentai.googleapis.com/v1/projects/{PROJECT_ID}/locations/{LOCATION}/processors/{PROCESSOR_ID}:process ↗ Unit price: Enterprise Document OCR: $1.50 per 1,000 pages (first 1,000 pages/month free); Form Parser and Custom extractor $30 per 1,000 pages; Layout Parser $10 per 1,000 pages ↗	Requires a GCP project and service-account/OAuth2 credentials (standard Google Cloud IAM auth); region-specific endpoints (global/us/eu); no official MCP server found.	—
Google Slides / Docs API	Read and programmatically edit Google Slides presentations (slides, shapes, text, tables) via REST API or MCP.	OAuth 2.0 `Authorization: Bearer <OAuth2 token>; scopes e.g. https://www.googleapis.com/auth/presentations`	official	free, no key (standard Google API usage quotas apply; requires GCP project + OAuth consent)	Base URL: https://slides.googleapis.com Free plan: no charge; subject to per-minute request quotas ↗ Rate limits: Reads 3,000/min per project, 600/min per user; writes 600/min per project, 60/min per user ↗ OAuth scopes: https://www.googleapis.com/auth/presentations, https://www.googleapis.com/auth/presentations.readonly, https://www.googleapis.com/auth/drive.file, https://www.googleapis.com/auth/drive.readonly ↗ Example call: read a presentation (slides, shapes, text): GET https://slides.googleapis.com/v1/presentations/{presentationId} ↗ MCP tools: 2: read_presentation, update_presentation ↗	Official remote MCP server (slidesmcp.googleapis.com, tools: read_presentation, update_presentation) is part of the Google Workspace Developer Preview Program - limited/early access. Non-OpenAPI Discovery Doc available at https://slides.googleapis.com/$discovery/rest?version=v1.	—
LlamaParse (LlamaCloud)	Parse, classify, extract, and split complex documents (tables, forms, charts) into LLM-ready structured data via API.	API key `Authorization: Bearer <LLAMA_CLOUD_API_KEY>`	official	Free tier: Free plan with 10K credits/month; beyond that credit-based at $1.25/1,000 credits (1-45+ credits/page depending on tier); 48h parse cache is free to re-query	Free plan: 10K credits (1,000 credits = $1.25), per month ↗ Rate limits: Free plan: 5 concurrent parse jobs, 5 concurrent extract jobs, 1 concurrent sheets job; paid tiers 20-100 ↗ Data policy: Parse cache retained 48 hours then permanently deleted; caching can be disabled ↗ Async jobs: REST API processes documents asynchronously as jobs ↗ Example call: start a parse job for an uploaded file: POST https://api.cloud.llamaindex.ai/api/v2/parse ↗ Unit price: per page: Fast 1 credit, Cost-effective 3, Agentic 10, Agentic Plus 45; $1.25 per 1,000 credits (Agentic = $0.0125/page) ↗	Regional MCP endpoints: mcp.llamaindex.ai (North America) and mcp.eu.llamaindex.ai (Europe), sign in with LlamaCloud account. Pricing tiers: Fast/Cost-effective/Agentic/Agentic Plus.	—
Mistral OCR API	Extract text, tables, and layout structure from PDFs and images as structured Markdown/JSON via REST API.	API key `Authorization: Bearer <MISTRAL_API_KEY>`	community	Free tier: Free plan includes $10/month in API credits; then $4/1,000 pages (OCR), $5/1,000 pages (Document AI); 50% off via Batch API	Base URL: https://api.mistral.ai Free plan: $10 in API credits (Free plan), per month ↗ Data policy: Model-training opt-out available on all plans including Free ↗ Example call: OCR a PDF to markdown: POST https://api.mistral.ai/v1/ocr ↗ Unit price: Mistral OCR 4: $4 per 1,000 pages ($2 per 1,000 via Batch API); Document AI $5 per 1,000 pages ↗ MCP tools: 1: ocr_local_file ↗	Current model: Mistral OCR 4; block/table/bounding-box extraction requires OCR 4+/2512+; no official Mistral-maintained MCP server, only community implementations.	GitHub 0★ D-Diaa/mistral-ocr-mcp (2026-09-24 16:54 UTC)
PDF.co API	Convert, merge, split, edit, fill forms, OCR, and extract data from PDFs via REST API.	API key `x-api-key: <API_KEY> header`	official	trial credits: 10,000 credits for 1 month, no credit card required	Base URL: https://api.pdf.co/v1 Free plan: trial, one-time ↗ Data policy: Temporary files stored 1 hour by default, then auto-removed ↗ Async jobs: set async=true to get a jobId; poll POST /v1/job/check until success/failed ↗ Example call: extract text from a PDF (OCR, layout kept): POST https://api.pdf.co/v1/pdf/convert/to/text ↗ Unit price: PDF to text: 21 credits per page (simple, no OCR: 4); Basic plan $9.99/mo ($8.99 billed annually) includes 16,500 credits, about $0.0006/credit (plans at pdf.co/pricing) ↗ MCP tools: 38: pdf_to_json, pdf_to_csv, pdf_to_text, pdf_to_xls, pdf_to_xlsx, pdf_to_xml … ↗	Base URL https://api.pdf.co/v1; credit-based usage pricing after trial; code samples (not formal SDKs) at github.com/pdfdotco/pdf-co-api-samples. Official MCP server: uvx pdfco-mcp (X_API_KEY env var).	—
Reducto API	Parse, extract, split, edit, and classify complex documents into structured JSON/Markdown via REST API or SDKs.	API key `Authorization: Bearer <REDUCTO_API_KEY>`	official	Trial: $150 in free credits (15,000 credits) one-time for new Standard accounts; then standard rates; Startup Program gives extra credits to qualifying teams	Base URL: https://platform.reducto.ai Free plan: 15,000 free credits (~$150) for Standard accounts; card requirement not stated on the pricing page, one-time ↗ Rate limits: Concurrent pages: Standard 200, Growth 350, Enterprise 500+ ↗ Data policy: Zero Data Retention agreement available on Growth and Enterprise tiers ↗ Async jobs: async parse with job polling or webhooks (sync /parse also available) ↗ Example call: parse a document into chunks: POST https://platform.reducto.ai/parse ↗ Unit price: Standard: Parse $10 per 1,000 pages; Extract $20 per 1,000 pages (parsing included) ↗ MCP tools: 6: parse_document, extract_data, split_document, classify_document, edit_document, upload_file ↗	Pay-as-you-go after free credits: ~$10/1,000 pages (Parse), ~$20/1,000 pages (Extract); official MCP repo at github.com/reductoai/mcp-server-reducto; hosted MCP variant does not support local file paths, only URLs.	—
SlideSpeak API	Generate, edit, and download PowerPoint/PDF presentations from text, documents, or outlines via REST API.	API key `X-API-Key: <key>`	official	Trial: free API trial, no credit card required; then API plans from $49/month (1,000 credits)	Base URL: https://api.slidespeak.co/api/v1 Free plan: trial, one-time, no card ↗ Data policy: API plans state no model training on your data ↗ Async jobs: POST /presentation/generate returns task_id; poll GET /task_status/{task_id}, then /presentation/download/{request_id} ↗ Example call: generate a presentation from text: POST https://api.slidespeak.co/api/v1/presentation/generate ↗ Unit price: 1 API credit per generated slide (1 credit = 5 outline slides); API plans from $49/mo for 1,000 credits (~$0.05 per slide) ↗	Async task-based workflow with polling (GET /task_status/{task_id}); download URLs are short-lived; no official language SDK, only a Postman collection plus the MCP server.	GitHub 14★ SlideSpeak/slidespeak-mcp (2026-09-24 16:54 UTC) SourceForge 1.0★/1 review (2026-09-24 16:57 UTC)
Unstructured API	Partition and extract structured elements (text, tables, layout) from PDFs, docs, and images for RAG pipelines via API.	API key `unstructured-api-key: <key> (verify exact header on api-reference page)`	official	Trial: 10,000 free pages one-time on signup, no credit card; then pay-as-you-go at $0.015/page	Base URL: https://transform.unstructured.io/api/v2 Free plan: 10,000 free pages at account start, then $0.015/page, one-time, no card ↗ Data policy: Zero Data Retention: processed data not stored permanently ↗ Async jobs: Prefer: wait=0 returns 202 with Location URL; poll job until completed/failed ↗ Example call: parse a document into elements (Transform Parse API): POST https://transform.unstructured.io/api/v2/parse ↗ Unit price: Pay-As-You-Go: $0.015 per page after the first 10,000 free pages ↗	Legacy /general/v0/general Partition endpoint is for prototyping single files; production use should use the Workflow API. Hosted Transform MCP parses documents; the UNS-MCP repo only manages connectors/workflows and is no longer maintained. Current parse endpoint: POST https://transform.unstructured.io/api/v2/parse with header unstructured-api-key	GitHub 43★ Unstructured-IO/UNS-MCP (2026-09-24 16:54 UTC)

Documents, OCR and presentations APIs and MCP servers for AI agents

Other categories