We maintain Arcmira, a YouTube transcript search API for timestamped quotes, speaker appearances, mentions and sponsor research. We would like to contribute a catalog page using Haystack's existing MCP adapter, and want to confirm the packaging requirement before opening a listing PR.
The repository README includes external technologies and asks for a PyPI or repository link. The integration guide also asks for an installable, runnable package. Our working connection uses the existing mcp-haystack package; it does not add a new Arcmira-specific component.
Would a page documenting that connection qualify, or must an Arcmira-specific package be published first? We are not asking for a model-provider designation or an untested research integration.
Proposed catalog entry
Runnable public-docs example
pip install haystack-ai==3.2.0 mcp-haystack==1.5.1
from haystack_integrations.tools.mcp import MCPToolset, StreamableHttpServerInfo
def main():
docs = MCPToolset(
server_info=StreamableHttpServerInfo(
url="https://arcmira.com/docs/mcp", max_retries=0
),
tool_names=["search_arcmira_api", "query_docs_filesystem_arcmira_api"],
)
try:
docs.warm_up()
search = next(tool for tool in docs if tool.name == "search_arcmira_api")
print(search.invoke(query="How do I authenticate and search YouTube transcripts?"))
finally:
docs.close()
if __name__ == "__main__":
main()We ran this exact example against the live documentation server on October 5, 2026 UTC. It returned 10 documentation content blocks with no tool error. Both installed top-level versions are the ones shown above. The connection exposes only documentation search and documentation file reads, requires no Arcmira key and does not make a model call. The filesystem tool reads hosted documentation, not local files.
The proposed page would clearly separate public documentation lookup from authenticated transcript research. Research needs an Arcmira account; paid reads use credits from the plan, then the configured on-demand budget. This test did not run authenticated research, paid operations or LLM inference.
The linked SDK repository is Apache-2.0. The hosted API and proprietary research MCP are separate products. We would include these distinctions, install instructions, the complete example and source/support links in the page. We have a local draft ready and can adapt it to the catalog's preferred scope.
We maintain Arcmira, a YouTube transcript search API for timestamped quotes, speaker appearances, mentions and sponsor research. We would like to contribute a catalog page using Haystack's existing MCP adapter, and want to confirm the packaging requirement before opening a listing PR.
The repository README includes external technologies and asks for a PyPI or repository link. The integration guide also asks for an installable, runnable package. Our working connection uses the existing
mcp-haystackpackage; it does not add a new Arcmira-specific component.Would a page documenting that connection qualify, or must an Arcmira-specific package be published first? We are not asking for a model-provider designation or an untested research integration.
Proposed catalog entry
Runnable public-docs example
We ran this exact example against the live documentation server on October 5, 2026 UTC. It returned 10 documentation content blocks with no tool error. Both installed top-level versions are the ones shown above. The connection exposes only documentation search and documentation file reads, requires no Arcmira key and does not make a model call. The filesystem tool reads hosted documentation, not local files.
The proposed page would clearly separate public documentation lookup from authenticated transcript research. Research needs an Arcmira account; paid reads use credits from the plan, then the configured on-demand budget. This test did not run authenticated research, paid operations or LLM inference.
The linked SDK repository is Apache-2.0. The hosted API and proprietary research MCP are separate products. We would include these distinctions, install instructions, the complete example and source/support links in the page. We have a local draft ready and can adapt it to the catalog's preferred scope.