Click on "Install 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., "@HunyuanVideo 1.5 720p MCP Using this image, generate a 5-second 720p video with paper texture preserved and slow line movement."
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.
RunPod Serverless에서 HunyuanVideo-1.5 720P-I2V를 실행하고 MCP로 비동기 작업을 제출·조회·취소하는 프로젝트입니다.
모델: HunyuanVideo-1.5 720P-I2V
출력: 1280×720, 24fps, 무음 H.264 MP4
길이: 5~10초
GPU: A40 48GB 우선, RTX A6000 48GB 대체, A100 PCIe 80GB 최종 fallback
Network Volume: qs5gpwp1p1, 60GB, CA-MTL-3
Worker: 최소 0, 최대 1, 유휴 5초 후 정지
Serverless 48GB Flex 계산 기준: 현재 $1.22/시간(실제 청구는 RunPod 가격 우선)
결과: 비공개 Cloudflare R2에 저장 후 최대 7일짜리 다운로드 URL 반환
에이전트 입력: MCP에서 로컬 PNG/JPEG 또는 base64/data URI를 R2에 직접 업로드
Related MCP server: runway-mcp
Claude/Cursor에서 이미지를 첨부한 뒤 다음처럼 요청할 수 있습니다.
이 이미지를 업로드하고, 종이 질감은 유지한 채 선이 천천히 움직이는 5초 720p 영상으로 만들어줘.
에이전트는 다음 순서로 MCP 도구를 호출합니다.
upload_video_input에 첨부 이미지의 local_path, image_data_base64, image_data_uri 중 하나를 전달
반환된 asset_id를 submit_video_job에 전달
get_video_job으로 완료 상태를 확인
get_video_result로 새 다운로드 링크를 받음
업로드는 PNG/JPEG, 최대 20MB, 한 변 64~8192px, 최대 40MP로 제한됩니다. 이미지는 실제로 디코딩해 손상 여부를 확인하고 콘텐츠 해시 기반 R2 키로 저장하므로 같은 파일은 같은 asset_id를 사용합니다. 16:9가 아닌 입력은 1280×720 workflow에서 잘리거나 재구성될 수 있다는 경고를 반환합니다.
첫 5초 GPU 테스트 비용 승인 (GPU-TEST)
테스트 성공 후 장면 8개 전체 생성 비용 승인 (ALL-8)
GitHub 로그인, Worker 빌드, R2 입력 8개 업로드, RunPod Template·Endpoint 생성은 완료되었습니다. 추가 API key 입력은 필요하지 않습니다.
.env는 Git에 포함되지 않습니다. API key와 R2 Secret은 채팅이나 저장소에 올리지 않습니다.
비용 표시는 Serverless Flex의 현재 48GB/A100 시간당 가격을 사용한 추정치이며, 실제 RunPod 청구가 우선합니다.
전체 절차는 READY_TO_DEPLOY.md에 정리되어 있습니다.
가장 쉬운 방법은 Finder에서 START_HERE.command를 더블클릭하는 것입니다. 완료된 배포 상태를 확인한 뒤 GPU 비용 승인을 안내하며, 승인 문구가 정확히 입력되기 전에는 유료 작업을 실행하지 않습니다.
python3 scripts/preflight.py
scripts/finalize_deployment.sh --approve-public-code-and-r2-upload
.venv/bin/python scripts/smoke_test.py --approve-gpu-costdeploy_runpod.py는 기본적으로 dry-run이며 --apply를 지정할 때만 템플릿과 Endpoint를 생성합니다.
Estimate runtime and compute cost without contacting RunPod.
| Name | Required | Description | Default |
|---|---|---|---|
| asset_id | No | scene-01 | |
| duration_sec | No |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description bears full responsibility for behavioral disclosure. It only states it performs estimation without contacting RunPod, but omits details like whether it's read-only, what data it accesses, or any side effects.
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 a single sentence with no wasted words, but it is too brief to convey necessary information. It earns its place but lacks substance.
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 two parameters with no schema descriptions, an output schema (unseen), and no annotations, the one-sentence description is insufficient for an agent to understand inputs or behavior. It is incomplete.
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 explain parameter meaning. However, the description does not mention asset_id or duration_sec at all, leaving their roles completely unclear.
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 function: estimating runtime and compute cost for a video job, and distinguishes from sibling tools like list_video_models or get_video_job by specifying the estimation nature and the 'without contacting RunPod' aspect.
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, nor does it mention any prerequisites or context for use. The sibling tools are listed but not discussed.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Return the current state and metadata for a Hunyuan job.
| Name | Required | Description | Default |
|---|---|---|---|
| job_id | Yes |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Description indicates a read-only operation ('Return'). No annotations provided, but the description is minimally transparent. Lacks details on rate limits or permissions, which are reasonable for a simple getter.
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?
One short sentence that directly states purpose. No wasted words, but no structural elements (e.g., sections). Acceptable for a simple tool.
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 one required parameter and the presence of an output schema, the description adequately covers the tool's purpose. Does not elaborate on return format or errors, but the output schema presumably handles that.
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?
With 0% schema description coverage, the description adds no extra meaning to the 'job_id' parameter beyond its name and type. It does not explain format, origin, or constraints of the job ID.
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?
Description clearly states the verb 'Return' and resource 'Hunyuan job', specifying the output 'current state and metadata'. This differentiates it from sibling tools like list_video_models or estimate_video_job.
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?
Usage is implied: retrieve job status by job_id. However, no explicit guidance when to use vs alternatives, such as when a specific job's details are needed after submission.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
List recent images uploaded through upload_video_input.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No |
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full burden. It only says 'list recent images', implying a read operation, but does not disclose any behavioral details such as sorting order, what 'recent' means, or whether there are any side effects. This is minimal disclosure for a listing 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 a single sentence, very concise. It is front-loaded with the verb. While succinct, it could be improved by adding a brief note about the parameter or return value without losing conciseness.
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 low complexity (1 optional param, list operation) and no annotations, the description is insufficient. It does not mention what the output contains (even though an output schema exists) nor the effect of the limit parameter. An agent would need to inspect the output schema to understand the return structure, but the description should provide at least a hint.
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 only parameter is 'limit' with a default value, but the description does not explain its meaning or usage. Since schema description coverage is 0%, the description should compensate, but it does not. The parameter name is self-explanatory to some extent, but no additional context is provided.
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 ('list') and the resource ('recent images uploaded through upload_video_input'), distinguishing from sibling tools that deal with models, estimates, and jobs. However, there is a minor inconsistency: the tool name says 'video_inputs' while the description says 'images', which could cause confusion.
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?
No guidance is provided on when to use this tool versus alternatives (e.g., list_video_models). The description only states what the tool does, leaving the agent to infer usage context without explicit when-to-use or when-not-to-use instructions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
List the single enabled HunyuanVideo generation profile.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The tool has no annotations, and the description does not disclose any behavioral traits such as side effects, permissions required, or rate limits. It is presumably a safe read-only operation, but this is not explicitly stated.
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 short and front-loaded, fitting in a single sentence. While concise, it borders on too minimal, but it still effectively communicates the core action.
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 parameters and an existing output schema, the description does not need to explain return values. However, it fails to provide context about the relationship to sibling tools or what the output represents, leaving room for improvement.
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 tool has zero parameters, so the description need not add parameter meaning. The schema covers all parameters, meeting the baseline score of 4 as specified.
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 tool lists the 'single enabled HunyuanVideo generation profile', which distinguishes it from sibling tools like 'list_uploaded_video_inputs' and 'estimate_video_job'. However, it does not explain what a 'generation profile' is, leaving some ambiguity.
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?
No guidance is provided on when to use this tool versus alternatives. The description does not mention any prerequisites, conditions, or context for invocation.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
v0.1.0estimate_video_jobget_video_joblist_uploaded_video_inputslist_video_modelsEach tool has a distinct purpose: listing models, listing uploaded inputs, estimating a job, and getting a job's state. There is no overlap or ambiguity.
All tools follow a consistent verb_noun pattern (list_, estimate_, get_), making it predictable for agents.
With 4 tools, the set is small but reasonable for a focused server. A few more tools might be expected, but the count is not problematic.
The tool surface has significant gaps: there is no tool to upload video inputs or to create a video generation job, which are core operations implied by the server's purpose.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
MCP server for Hailuo (MiniMax) AI video generation
Create and manage AI image and video generations through Quriov's fixed public MCP tools.
Plan, compare, price, generate, and recover AI video from compatible MCP clients.
Run multi-step AI pipelines for video, image, audio and text: upload media, run, poll results.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/ssy2205/hunyuan-video-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
