search_llamagen_docsSearch the integration and Comic API documentation.
Agent-readable install guide
This page describes the capability that exists now: a hosted Comic MCP protected by a LlamaGen API token. It does not claim that a public marketplace plugin or OAuth installation has shipped.
Follow this order every time. The agent should not spend credits or begin a comic run until the account passes the balance check.
Open LlamaGen Settings → API, click Create Key, then use the copy button beside the new key. Keep the token private.
Open API settingsUse https://llamagen.ai/api/mcp with Streamable HTTP and the Authorization header: Bearer <API_TOKEN>.
The agent must call get_comic_api_usage. If remaining API credits are below 5,000, it must stop before reading the creation docs or generating.
If the balance is below 5,000, upgrade the Comic API plan or buy a one-time API credit pack.
View Comic API pricingWhen at least 5,000 credits remain, read the current Comic API documentation, propose the comic plan, and wait for approval before generation.
Read Comic API docsUse the live tool name: get_comic_api_usage
Older screenshots may showget_api_usage. The current MCP manifest exposesget_comic_api_usage, so the agent must call that exact tool.
Create or rotate a token in LlamaGen Settings → API. Keep it in your shell environment; do not paste the secret into a chat or repository.
Run the command below in a terminal. It registers the hosted Streamable HTTP MCP for your user scope.
Run `claude mcp get llamagen`, start a new Claude Code session if needed, then use `/mcp` to confirm the server and tools.
Give Claude access only to the scripts and references in scope, then approve the proposed batch before generation.
1 · Set the token
Keep the real value out of chat history and source control
export LLAMAGEN_API_TOKEN='replace-with-your-token'2 · Register LlamaGen
Claude Code remote HTTP MCP
claude mcp add --scope user --transport http llamagen https://llamagen.ai/api/mcp --header "Authorization: Bearer ${LLAMAGEN_API_TOKEN}"3 · Verify
Confirm the connection before starting a batch
claude mcp get llamagen
# In Claude Code: /mcpThe public MCP manifest and server card are the source of truth. If this list differs in the host, stop and reconnect before generating.
search_llamagen_docsSearch the integration and Comic API documentation.
get_comic_api_usageCheck the plan and remaining API credits before reading the creation docs or generating.
create_comic_generationStart a comic, manga, webtoon, or storyboard generation.
get_comic_generation_statusRead task state, generated panels, assets, and errors.
Ask for a page or batch plan first. Creation consumes LlamaGen API credits.
Claude Code usage pays for reasoning; LlamaGen credits pay for comic generation.
Use the host to control the workflow and LlamaGen to perfect layout, bubbles, assets, and delivery.
Working prompt
Read https://llamagen.ai/claude and use the connected LlamaGen MCP.
Required preflight:
1. Ask me to open https://llamagen.ai/settings?tab=api, click Create Key, copy the API token, and store it as LLAMAGEN_API_TOKEN without pasting it into chat or committing it.
2. Connect https://llamagen.ai/api/mcp with Authorization: Bearer <API_TOKEN>, then list the four expected LlamaGen tools.
3. Call get_comic_api_usage before reading the creation docs or generating. Report LlamaGen API credits separately from Claude usage.
4. If remaining API credits are below 5,000, stop and send me to https://llamagen.ai/comic-api/pricing to upgrade my plan or buy an API credit pack.
5. When the balance is at least 5,000 credits, read https://llamagen.ai/comic-api/docs to learn the current comic-creation flow.
6. Read only the scripts, references, or production notes I place in scope.
7. Propose a page, scene, or batch plan and wait for approval.
8. Use create_comic_generation only after approval, then poll get_comic_generation_status.
9. Return generation IDs and assets, and explain which visual refinements belong in the LlamaGen editor.