Server data from the Official MCP Registry
Drive Instavar Studio video jobs (create, render, approve, publish, metrics) with an API key.
About
Drive Instavar Studio video jobs (create, render, approve, publish, metrics) with an API key.
Security Report
This is a well-structured MCP server for the Instavar video API with strong security fundamentals. Authentication is properly enforced via environment variables, permissions are narrowly scoped to a single HTTP API endpoint, and dangerous operations are avoided. The code includes thoughtful prompt-injection defenses via nonce-delimited data envelopes. Minor code quality observations exist but do not materially impact security. Supply chain analysis found 2 known vulnerabilities in dependencies (1 critical, 0 high severity). Package verification found 1 issue.
7 files analyzed · 6 issues found
Security scores are indicators to help you make informed decisions, not guarantees. Always review permissions before connecting any MCP server.
Permissions Required
This plugin requests these system permissions. Most are normal for its category.
What You'll Need
Set these up before or after installing:
Environment variable: INSTAVAR_API_KEY
Environment variable: INSTAVAR_BASE_URL
How to Install
Add this to your MCP configuration file:
{
"mcpServers": {
"com-instavar-mcp-server": {
"env": {
"INSTAVAR_API_KEY": "your-instavar-api-key-here",
"INSTAVAR_BASE_URL": "your-instavar-base-url-here"
},
"args": [
"-y",
"@instavar/mcp-server"
],
"command": "npx"
}
}
}Documentation
View on GitHubFrom the project's GitHub README.
@instavar/mcp-server
Make local videos with Instavar's open-source Remotion templates from Claude Code or Codex. The public template, schema, scaffold, and troubleshooting tools work without an account or API key.
Add an Instavar API key when you want the hosted workflow: create → render → approve → publish → metrics. The authenticated tools are thin HTTP clients to the hosted Instavar API. The package holds no database connection and stores no data of its own.
Which model runs?
The MCP host supplies the general-purpose LLM.
- In Claude Code, the selected Claude model interprets the request and calls Instavar tools.
- In Codex, the selected OpenAI model interprets the request and calls the same tools.
- Other compatible hosts use their own selected model and authentication.
The MCP server does not provide Sonnet, GPT, or another general-purpose model. It exposes typed Instavar capabilities to the model already running in the user's host. Depending on the host and sign-in method, that model usage may draw from a personal or workspace subscription, separate usage credits, or API billing. Those provider rules can change.
Local Remotion rendering uses your own machine. Instavar meters or charges its hosted services, including rendering, storage, specialist media generation, QA, publishing, and other declared job costs. Using a subscription-backed agent does not make those services free.
For the full architecture, see
docs/architecture/studio-agent-workflow-architecture.md.
For dated provider evidence, see
reports/research/subscription-backed-agent-runtime-2026-07-10.md.
Setup
Add the server to your MCP client config:
{
"mcpServers": {
"instavar": {
"command": "npx",
"args": ["-y", "@instavar/mcp-server"]
}
}
}
This starts with the free local template tools. To unlock hosted Studio tools, create an API key at https://instavar.com/studio/settings, then add it:
{
"mcpServers": {
"instavar": {
"command": "npx",
"args": ["-y", "@instavar/mcp-server"],
"env": { "INSTAVAR_API_KEY": "ik_live_..." }
}
}
}
Restart the client. The free tools appear in both modes. Hosted tools appear when the key is configured.
Environment
| Var | Required | Default | Purpose |
|---|---|---|---|
INSTAVAR_API_KEY | no | — | Unlocks hosted Studio tools. Scoped to your org; revocable. |
INSTAVAR_BASE_URL | no | https://instavar.com | Override the API host. |
Tools
Free local tools:
| Tool | What it does |
|---|---|
get_capabilities | Reports free and optional account features. |
list_templates | Lists published Remotion template families. |
get_template | Returns one template contract. |
get_video_schema | Returns the public VideoSpec JSON Schema. |
get_example | Returns a starter VideoSpec. |
scaffold_video_spec | Creates a deterministic VideoSpec skeleton. |
get_render_instructions | Returns the local validation and render workflow. |
diagnose_render_error | Classifies common local render failures. |
Hosted tools, available when INSTAVAR_API_KEY is configured:
| Tool | What it does |
|---|---|
list_jobs | List recent jobs (newest first). |
get_job_status | Runs, artifacts, verifications, current video + thumbnail URLs. |
get_video_state | Materialized composition state for a job. |
get_job_metrics | Platform engagement snapshots for a published job. |
get_cost_summary | Infra cost summary (Lambda / RunPod / R2 / WaveSpeed / PoYo). |
inspect_workflow | Read the current workflow and its version history. |
create_video_brief | Create a job from a structured brief and start rendering. |
edit_video_brief | Patch a brief (incl. objective / publishTarget); re-renders unless only publishTarget changed. |
retry_video_qa | Retry degraded or unavailable QA against the same current frozen video artifact. |
audit_video_qa | Run an advisory Qwen3.6-Plus second opinion against the same frozen video. |
regenerate_narration | Regenerate narration and its dependent video from the unchanged durable brief. |
draft_workflow | Store a complete workflow using the inspected expectedVersion (0 only when none exists). |
patch_workflow | Create a conflict-aware workflow version from a recursive patch. |
validate_workflow | Validate and compile the inspected workflow version without rendering. |
preview_workflow | Read the compiled UniversalVideo brief without rendering. |
compare_workflows | Compare two durable workflow versions field by field. |
approve_workflow | Approve the inspected valid version without approving or publishing the video job. |
approve_job | Approve a rendered job so it can be published. |
publish_job | Publish an approved job to its connected social destination. |
connect_account | Start connecting a social account (headless OAuth pairing). |
connect_account_status | Poll a pending account connection until it is connected. |
Workflow inspection, preview, and comparison require read. Workflow draft,
patch, validation, and approval require write. publish_job requires
publish.
Workflow tools reject unknown arguments instead of silently discarding them.
API calls time out after 15 seconds. A stale-version 409 tells the caller to
inspect the latest version, while an unavailable upload or other dependency
conflict tells the caller to resolve that dependency before retrying.
Connecting a social account from the terminal
connect_account returns an approve URL and a short confirm code. Open
the URL in a browser where you are signed in to Instavar, check the code matches
the one in your terminal, and approve — you go through the platform's normal
sign-in, then return to Instavar. Poll connect_account_status with the returned
pairingId until it reports connected. Pairings expire after 10 minutes and
can only be approved by an owner/admin of the workspace.
CLI
npx @instavar/mcp-server --version
npx @instavar/mcp-server --help
License
MIT
Reviews
No reviews yet
Be the first to review this server!
More Developer Tools MCP Servers
Git
Freeby Modelcontextprotocol · Developer Tools
Read, search, and manipulate Git repositories programmatically
Fetch
Freeby Modelcontextprotocol · Developer Tools
Web content fetching and conversion for efficient LLM usage
Toleno
Freeby Toleno · Developer Tools
Toleno Network MCP Server — Manage your Toleno mining account with Claude AI using natural language.
mcp-creator-python
Freeby mcp-marketplace · Developer Tools
Create, build, and publish Python MCP servers to PyPI — conversationally.
MarkItDown
Freeby Microsoft · Content & Media
Convert files (PDF, Word, Excel, images, audio) to Markdown for LLM consumption
MCP Marketplace
Freeby mcp-marketplace · Developer Tools
Search and install MCP servers from inside your AI client.
