2slides MCP Server
OfficialThis MCP server provides programmatic access to 2slides.com's presentation generation API through Claude Desktop.
Generate presentation slides - Create slides by providing a theme ID, content description, and preferred language. Supports both synchronous (wait for immediate results) and asynchronous (job-based) generation modes.
Search for presentation themes - Find suitable slide templates using keyword queries (e.g., "8 stages") with optional result limits (up to 100 themes).
Monitor slide generation jobs - Poll the status of asynchronous jobs using a job ID (recommended every 10 seconds) until completion, then retrieve the download URL for the generated slide deck.
Provides tools for generating presentation slides through the 2slides.com API, including searching themes, creating slide content from user input, and monitoring job status for slide generation tasks
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@2slides MCP Serversearch for business presentation themes"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
2slides MCP Server

Expose 2slides.com tools to MCP clients: generate PowerPoint decks and custom-designed slides, add AI voice narration, and get the finished files back as download links.
Get Your API Key
Before connecting, get an API key from 2slides.com/api.
Connect
2slides runs a hosted Streamable HTTP MCP endpoint, and also ships this stdio package. Both expose the same 7 tools against the same account and the same credits.
Streamable HTTP | stdio (this package) | |
Setup | one command, nothing to install | needs Node.js 18+, npx fetches the package |
Updates | server-side, always current | re-downloaded on your machine |
Runs in | any remote-capable MCP client | any client that can spawn a local process |
Option 1 — Streamable HTTP (recommended)
Anthropic's MCP docs call HTTP "the recommended option for connecting to remote MCP servers", and 2slides is a hosted service with no local files or processes to reach, so start here.
Claude Code
claude mcp add --transport http 2slides "https://2slides.com/api/mcp?apikey=YOUR_2SLIDES_API_KEY"Verify with claude mcp list — the server should report ✔ Connected.
Project config (.mcp.json, shared with your team)
{
"mcpServers": {
"2slides": {
"type": "http",
"url": "https://2slides.com/api/mcp?apikey=YOUR_2SLIDES_API_KEY"
}
}
}Claude Desktop / claude.ai: add it as a custom connector at claude.ai/customize/connectors with the same URL.
Option 2 — stdio (local process)
Use this when the client can only spawn local processes, when the machine has no direct route to 2slides.com, or when you want to pin and patch the server yourself.
Claude Code
claude mcp add 2slides -e API_KEY=YOUR_2SLIDES_API_KEY -- npx -y 2slides-mcp@latestClaude Desktop — edit ~/Library/Application Support/Claude/claude_desktop_config.json
(%APPDATA%\Claude\claude_desktop_config.json on Windows):
{
"mcpServers": {
"2slides": {
"command": "npx",
"args": ["-y", "2slides-mcp@latest"],
"env": {
"API_KEY": "YOUR_2SLIDES_API_KEY"
}
}
}
}The -y flag lets npx install the package without prompting. Fully restart Claude Desktop
afterwards; in a chat, open the tools panel and you should see the tools below.
Available Tools
slides_generate (POST /api/v1/slides/generate)
Generate a PowerPoint presentation using a pre-built theme (Fast PPT). Use themes_search first to find a themeId.
Args:
userInput(string, required) — presentation contentthemeId(string, required) — theme ID fromthemes_searchresponseLanguage(string, optional, default:Auto) — output language; supported values:Auto,English,Spanish,Arabic,Portuguese,Indonesian,Japanese,Russian,Hindi,French,German,Greek,Vietnamese,Turkish,Thai,Polish,Italian,Korean,Simplified Chinese,Traditional Chinesemode(optional, default:sync) —syncreturnsdownloadUrlimmediately;asyncreturns ajobIdto poll withjobs_get
Credits: 1 per page
Example:
{ "themeId": "st-1756528793701-fcg5fblt2", "userInput": "Generate a 5-slide intro to machine learning", "responseLanguage": "English", "mode": "sync" }
themes_search (GET /api/v1/themes/search)
Search pre-built themes to use with slides_generate.
Args:
query(string, required) — search keyword (e.g.business,minimal,dark)limit(number, optional, default:10, range:1–100)
Returns: array of themes with
id,name,description,previewUrl,tagsCredits: free
Example:
{ "query": "business", "limit": 5 }
jobs_get (GET /api/v1/jobs/{jobId})
Poll a job's status and retrieve the result when ready.
Args:
jobId(string, required)
Returns:
jobId,status(pending/processing/success/failed),message,downloadUrl(when success),slidePageCount,createdAt,updatedAt,duration(ms),jobUrl(Nano Banana jobs only)Credits: free
Note: Poll every 20 seconds until
statusissuccessorfailed.Example:
{ "jobId": "D8h9VYDGdTlZ6wWSEoctF" }
slides_create_like_this (POST /api/v1/slides/create-like-this)
Generate slides (Nano Banana) that match the style of a reference image.
Args:
userInput(string, required) — presentation contentreferenceImageUrl(string, required) — URL of the style reference imageresponseLanguage(string, optional, default:Auto)aspectRatio(string, optional, default:16:9) — valid values:1:1,2:3,3:2,3:4,4:3,4:5,5:4,9:16,16:9,21:9resolution(string, optional, default:2K) —1K,2K,4Kpage(number, optional, default:0) —0= auto-detect,1–100= exact page countcontentDetail(string, optional, default:standard) —conciseorstandardmode(string, optional, default:async) —asyncreturnsjobId;syncreturnsdownloadUrlimmediately
Credits: 100 per page (1K/2K) or 200 per page (4K)
Example:
{ "userInput": "Create a presentation about AI trends", "referenceImageUrl": "https://example.com/style-reference.jpg", "aspectRatio": "16:9", "resolution": "2K", "page": 5, "contentDetail": "standard", "mode": "async" }
slides_create_pdf_slides (POST /api/v1/slides/create-pdf-slides)
Generate custom-designed slides (Nano Banana) from text content without a reference image.
Args:
userInput(string, required) — presentation contentdesignStyle(string, optional) — free-text description of the visual style; leave empty for default stylingresponseLanguage(string, optional, default:Auto)aspectRatio(string, optional, default:16:9) — valid values:1:1,2:3,3:2,3:4,4:3,4:5,5:4,9:16,16:9,21:9resolution(string, optional, default:2K) —1K,2K,4Kpage(number, optional, default:0) —0= auto-detect,1–100= exact page countcontentDetail(string, optional, default:standard) —conciseorstandardmode(string, optional, default:async) —asyncreturnsjobId;syncreturnsdownloadUrlimmediately
Credits: 100 per page (1K/2K) or 200 per page (4K)
Example:
{ "userInput": "Quarterly business review for Q3 2025", "designStyle": "modern, dark background, bold typography", "aspectRatio": "16:9", "resolution": "2K", "page": 8, "mode": "async" }
slides_generate_narration (POST /api/v1/slides/generate-narration)
Add AI voice narration to a completed Nano Banana job (slides_create_like_this or slides_create_pdf_slides only; Fast PPT not supported). Always async — poll result with jobs_get.
Args (single speaker,
mode: "single"default):jobId(string, required) — UUID of a completed Nano Banana jobmode(string, optional, default:single) —singleormultispeakerName(string, optional) — narrator namevoice(string, optional) — one of 30 voices (see list below)contentMode(string, optional, default:standard) —conciseorstandardincludeIntro(boolean, optional, default:true)
Args (two speakers,
mode: "multi"):jobId(string, required)mode(string, required) —multispeaker1Name(string, required)speaker2Name(string, required)speaker1Voice(string, optional) — voice for speaker 1speaker2Voice(string, optional) — voice for speaker 2contentMode(string, optional, default:standard) —conciseorstandard
Supported voices (30):
Puck,Aoede,Charon,Kore,Fenrir,Zephyr,Leda,Orus,Callirrhoe,Autonoe,Enceladus,Iapetus,Umbriel,Algieba,Despina,Erinome,Algenib,Rasalgethi,Laomedeia,Achernar,Alnilam,Schedar,Gacrux,Pulcherrima,Achird,Zubenelgenubi,Vindemiatrix,Sadachbia,Sadaltager,SulafatCredits: 210 per page (10 text + 200 audio)
Example:
{ "jobId": "a1b2c3d4-e5f6-7890-abcd-ef1234567890", "mode": "multi", "speaker1Name": "Alice", "speaker2Name": "Bob", "speaker1Voice": "Aoede", "speaker2Voice": "Puck", "contentMode": "standard" }
slides_download_slides_pages_voices (POST /api/v1/slides/download-slides-pages-voices)
Export slide pages and audio narration as a ZIP file. The job must be completed and have narration generated.
Args:
jobId(string, required) — UUID of a job with completed narration
Returns:
downloadUrl(expires in 1 hour),fileName,expiresIn(3600 seconds)ZIP contents:
pages/ page_01.png, page_02.png, … voices/ page_01.wav, page_02.wav, … transcript.txtCredits: free
Example:
{ "jobId": "a1b2c3d4-e5f6-7890-abcd-ef1234567890" }
Typical Workflows
Fast PPT (theme-based PowerPoint):
themes_search→ get athemeIdslides_generate(sync) → getdownloadUrldirectly
Nano Banana (custom-designed slides):
slides_create_like_thisorslides_create_pdf_slides(async) → getjobIdjobs_get(poll every 20s) → getdownloadUrlwhenstatus: "success"
Nano Banana with narration:
slides_create_like_thisorslides_create_pdf_slides→ getjobIdjobs_get→ wait for slides to completeslides_generate_narration→ get narrationjobIdjobs_get→ wait for narration to completeslides_download_slides_pages_voices→ get ZIP download link
Troubleshooting (Claude Desktop)
If tools don't appear in Claude, verify the config path is absolute and restart the app.
Check Claude MCP logs:
tail -n 50 -f ~/Library/Logs/Claude/mcp*.logFor stdio MCP servers, avoid logging to stdout; this server only logs errors to stderr.
References
Build an MCP server (official docs): https://modelcontextprotocol.io/docs/develop/build-server
2slides API docs: https://2slides.com/api
2slides: https://2slides.com
2slides Templates: https://2slides.com/templates
Available Tools
3 toolsjobs_getA
Get job status/result by jobId from 2slides. Please check every 20 seconds until the status is success.
| Name | Required | Description | Default |
|---|---|---|---|
| jobId | Yes |
TDQS
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden. It discloses the polling behavior ('check every 20 seconds') and expected outcome ('until status is success'), which are valuable behavioral traits. However, it lacks details on error handling, timeout limits, or what happens if the job fails, leaving gaps in transparency.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is appropriately sized and front-loaded, with the core purpose stated first and usage guidance following. Both sentences earn their place by providing essential information without redundancy, making it efficient and well-structured.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no annotations, no output schema, and a simple input schema, the description is moderately complete. It covers the tool's purpose and polling behavior but lacks details on response format, error cases, or integration with sibling tools, leaving room for improvement in guiding an agent fully.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The description adds meaning by specifying that the jobId parameter is used to retrieve status/result from 2slides, which clarifies its purpose beyond the schema's basic string validation. With 0% schema description coverage and only one parameter, this compensation is adequate, though it could detail format or source of jobId.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb ('Get') and resource ('job status/result'), specifying it retrieves information by jobId from 2slides. However, it doesn't explicitly differentiate from sibling tools like slides_generate or themes_search, which appear to be different operations rather than alternative status-checking methods.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides implied usage guidance by mentioning 'Please check every 20 seconds until the status is success,' suggesting this tool is for polling job completion. However, it doesn't explicitly state when to use this tool versus alternatives or mention prerequisites like needing a valid jobId from another operation.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
slides_generateB
Generate slides with 2slides. Returns job info including jobId and downloadUrl when ready. Optional 'mode' can be 'sync' (default) or 'async'.
| Name | Required | Description | Default |
|---|---|---|---|
| themeId | Yes | ||
| userInput | Yes | ||
| responseLanguage | Yes | ||
| mode | No |
TDQS
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden. It discloses that the tool returns job info (jobId and downloadUrl) and mentions sync/async modes, adding some behavioral context. However, it lacks details on permissions, rate limits, error handling, or what '2slides' entails, leaving gaps in transparency for a generation tool.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is front-loaded with the main purpose and key return info, followed by an optional parameter note. It's concise with two sentences, but could be slightly improved by integrating the mode explanation more seamlessly or adding brief context for other parameters.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no annotations, no output schema, and 4 parameters with 0% schema coverage, the description is incomplete. It covers the tool's purpose and return values partially, but misses details on parameter meanings, error cases, and full behavioral traits, making it only minimally viable for a generation tool with multiple inputs.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 0%, so the description must compensate. It only explains the 'mode' parameter with enum values, but doesn't add meaning for 'themeId', 'userInput', or 'responseLanguage'. This leaves three out of four parameters without semantic clarification, failing to adequately compensate for the low schema coverage.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the action ('Generate slides with 2slides') and resource ('slides'), specifying it's a generation tool. However, it doesn't differentiate from sibling tools like 'jobs_get' or 'themes_search', which are retrieval tools rather than creation tools, so it misses explicit sibling distinction.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage through the mention of 'mode' options ('sync' or 'async'), suggesting when to use based on timing preferences, but it doesn't provide explicit guidance on when to use this tool versus alternatives like 'jobs_get' for checking job status or 'themes_search' for finding themes, nor does it mention prerequisites or exclusions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
themes_searchC
Search 2slides themes by query. Optional limit (max 100).
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes | ||
| limit | No |
TDQS
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It mentions the search functionality and a limit constraint, but lacks critical behavioral details such as authentication requirements, rate limits, pagination behavior, or what the search returns (e.g., format, fields). For a search tool with no annotations, this is insufficient.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is extremely concise with two short sentences that directly convey the tool's function and key parameter constraint. Every word earns its place, and it's front-loaded with the core purpose. No wasted verbiage or redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no annotations, 0% schema description coverage, and no output schema, the description is incomplete. It covers basic purpose and a parameter limit but misses essential context like return format, error handling, or integration with sibling tools. For a search tool, this leaves significant gaps for an AI agent.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 0%, so the description must compensate. It adds meaning by specifying that 'query' is for searching themes and 'limit' is optional with a max of 100, which clarifies beyond the schema's basic types and constraints. However, it doesn't detail query syntax or examples, leaving some semantic gaps.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb ('Search') and resource ('2slides themes'), making the purpose evident. It specifies searching 'by query' which adds context. However, it doesn't distinguish this search tool from potential sibling search tools (though none are listed among siblings), keeping it from a perfect score.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool versus alternatives. It mentions an optional limit parameter but doesn't explain scenarios where this tool is appropriate or when other tools like 'slides_generate' might be preferred. No explicit usage context or exclusions are provided.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
3 tool updates
v1.0.0- Changed
jobs_get2 fields changed- added
Input schema / $schemaAdded value: +"http://json-schema.org/draft-07/schema#" - added
Input schema / additionalPropertiesAdded value: +false
- Changed
slides_generate3 fields changed- added
Input schema / $schemaAdded value: +"http://json-schema.org/draft-07/schema#" - added
Input schema / additionalPropertiesAdded value: +false - added
Input schema / properties / modeAdded value: +{ + "enum": [ + "sync", + "async" + ], + "type": "string" +}
- Changed
themes_search2 fields changed- added
Input schema / $schemaAdded value: +"http://json-schema.org/draft-07/schema#" - added
Input schema / additionalPropertiesAdded value: +false
3 tool updates
- First observed
jobs_get - First observed
slides_generate - First observed
themes_search
TDQS
Scored across 3 tools
Each tool has a clearly distinct purpose: jobs_get retrieves job status/results, slides_generate creates slides, and themes_search finds themes. There is no overlap in functionality, making it easy for an agent to select the correct tool without confusion.
All tool names follow a consistent verb_noun pattern (jobs_get, slides_generate, themes_search) with clear actions and targets. This predictability enhances usability and reduces cognitive load for agents.
With 3 tools, the server is well-scoped for slide generation and theme management, though it feels slightly thin. Each tool earns its place, but additional utilities like theme details or slide editing might be expected in a broader implementation.
The tools cover core workflows: generating slides, checking job status, and searching themes. Minor gaps exist, such as no direct theme retrieval or slide modification tools, but agents can work around these with the provided operations.
Maintenance
Related MCP Connectors
Create polished slide decks from text or YouTube links in seconds. Fetch video transcripts to tran…
Presentations.AI MCP server — create designed slide decks from a topic, text, or document.
Generate professional PowerPoint presentations from text, YouTube videos, or structured JSON data.…
Generate polished PowerPoint presentations from text prompts, YouTube videos, or structured outlin…
Related MCP Servers
- AlicenseAqualityDmaintenanceCreates and manipulates PowerPoint presentations with capabilities for adding various slide types, generating images, and incorporating tables and charts through natural language commands.11144MIT
- AlicenseBqualityDmaintenanceEnables users to create, edit, and manage PowerPoint presentations across Windows, macOS, and Linux through natural language via Claude Desktop. It provides comprehensive automation for slide management, text manipulation, and presentation styling using platform-specific adapters.131MIT
- FlicenseNot gradedqualityDmaintenanceEnables AI-powered generation of PowerPoint presentations with dynamic content, themes, and tones. Supports custom topics, audience, and slide count via the generate_ppt tool.-
- FlicenseNot gradedqualityDmaintenanceCreates AI-powered slide decks directly from Claude Desktop with live progress and QA scores.-