The tool surface · Adako docs

Why a router

A flat list of 231 tools costs thousands of tokens on every request, and some clients truncate a list that long. So the client sees a short, stable list: the tools people ask for by name, plus one router per platform. Everything else is one hop away and costs nothing to discover.

Callable by name

Tool	What it does
approve_proposal	Approve and execute a pending proposal
get_connections_status	Connection health per platform
get_tool_schema	Get the exact arguments a tool takes
get_usage_status	Plan and task usage
list_connected_accounts	List ad accounts (active and inactive)
list_pending_proposals	List proposals waiting for approval
reject_proposal	Reject a pending proposal
search_tools	Find the right Adako tool for a job
start_here	Start here — personalised next step
switch_primary_account	Switch the primary ad account for a platform
google_get_campaign_performance	Google Ads campaign performance with period-over-period change
meta_get_campaign_performance	Meta campaign performance
chatgpt_get_performance	Read ChatGPT Ads performance
tiktok_get_campaign_performance	Read TikTok campaign performance
linkedin_get_campaign_performance	LinkedIn campaign performance

Routers

Router	Covers	Tools
google_ads	Google Ads	69
meta_ads	Meta Ads	42
chatgpt_ads	ChatGPT Ads	22
tiktok_ads	TikTok Ads	29
linkedin_ads	LinkedIn Ads	33
monitoring	Monitoring and reporting	13
diagnostics	Diagnostics	8

The three actions

Every router takes the same shape.

router actions

google_ads(action="list_tools")
google_ads(action="get_tool_schema", tool_name="google_analyze_search_terms")
google_ads(action="execute", tool_name="google_analyze_search_terms",
           arguments={"customer_id": "1234567890", "date_range": "last_30_days"})

list_tools returns every tool in that router with a one-line summary, grouped by category. Free, and no task is charged.
get_tool_schema returns the JSON schema, required fields, enums and example prompts for one tool. Free.
execute runs the tool through the same pipeline as everything else: auth, validation, quota, the proposal gate, execution, read-back, audit. It bills the tool’s own cost.

A routed tool is not callable by name

Calling google_analyze_search_terms directly returns tool-not-found with near matches. The router is the address.

Finding the right tool

search_tools ranks the whole registry against the user’s own wording, across every platform, and returns the exact line to call each match with. It is free and instant, and it is the right first call whenever the tool is not obvious.

search, then call

search_tools(query="stop a campaign that is wasting money", top_k=5)
get_tool_schema(tool_names=["meta_pause_entity"])
meta_ads(action="execute", tool_name="meta_pause_entity",
         arguments={"entity_type": "campaign", "entity_id": "23851..."})

Both discovery tools cost nothing, so an assistant never has to guess a name. An unknown name returns suggestions rather than a result.

One read across several accounts

A platform router takes an accounts argument on read tools. It runs the tool once per account, one after another, and returns a single consolidated answer.

fan-out

google_ads(action="execute", tool_name="google_get_campaign_performance",
           arguments={"date_range": "last_30_days"}, accounts="all_active")

accounts takes a list of account ids or the literal "all_active".
Reads only. A write with accounts is refused; writes stay one account at a time.
Each account is billed as its own task. At most 20 accounts in one call.
Results are grouped per account, with the currency each one reports in.

Over REST

Routers and search_tools are MCP only. Over the REST API the URL is the dispatch: POST /api/v1/tools/{tool_name}/execute. Same pipeline, same quota, same proposals.