Find ranking gains
Positive rank changes with dates. Give your connected LLM the build instructions to create this reusable tool.
SE RankingTool specification · Live providers not tested
- What you provide
- Two domains, a country and a maximum result count.
- What you get
- Positive rank changes with dates.
How your tool works.
Join a supplied previous export to current rankings; enrich phrases; calculate position gains.
Your agent writes and saves the Python tool. Recurse supplies eligible API access and cloud execution. You can ask your agent to edit it and run it again.
The output.
Positive rank changes with dates.
The APIs in your tool.
Use eligible APIs through your Recurse balance. You do not need to open a separate provider account or fund a separate wallet for these operations. Operations requiring your own connected account remain subject to availability.
- SE Ranking$0.0179 / call
seranking.google.domain.ranked_keywordsPrice evidence and conditions
per_call · 100 credit per 1 call. Checked 2026-07-28. Source: docs. Confidence: documented.
100 credits per REQUEST, 0 per record. Because rows are free, pull `limit: 1000` per page rather than paging in small chunks — small pages multiply the per-request charge.
Price source · API documentation - SE Ranking$0.00179 / keyword
seranking.google.keywords.volumePrice evidence and conditions
per_result · 10 credit per 1 keyword. Checked 2026-07-28. Source: docs. Confidence: documented.
10 credits per keyword RETURNED, 0 per request. Batching 5,000 keywords into one call costs the same as 5,000 single-keyword calls (50,000 credits) but saves the round trips — batch for latency, not for price.
Price source · API documentation
These are dated API unit rates, not a whole-task quote. Your agent checks current availability and costs before execution. Settled provider costs pass through at 1×; preparation and execution compute are charged at 4×.
Build it with your agent.
Your agent asks for missing inputs and a total spending cap before chargeable work.
Read the full build instructions
Build and save a reusable Recurse tool called find_ranking_gains. Inputs: Two domains, a country and a maximum result count. Execution logic: Join a supplied previous export to current rankings; enrich phrases; calculate position gains. Output contract: Positive rank changes with dates. Declared APIs: seranking.google.domain.ranked_keywords, seranking.google.keywords.volume. Inspect these APIs' current executable status, native parameters, response examples and costs. Resolve and test response-field mappings before writing dependent steps; supplier response schemas in this research snapshot are incomplete. Do not silently substitute APIs. Implement the sequence, joins, filters and calculations in plain Python with from recurse_apis import call. Validate input and JSON output schemas in tools.yaml; lock dependencies in uv.lock. Keep interpretation outside the tool. Use at most 20 sequential API requests, 64 KiB inputs, 256 KiB results and 300 seconds. Limit fetched pages/posts and truncate evidence explicitly. Return null with a reason for missing fields; do not invent dates, contacts, prices or counts. Create success, empty, malformed-response and provider-error tests. Ask me for the missing inputs and one total spending cap before chargeable preparation. Build the standard source archive and perform Recurse's scoped HTTP upload; never send inline source to private_stage. Inspect the prepared function before private_call. Reuse prepared source for later invocations and retain its version ID. If this interface cannot build files and perform the upload, explain the missing capability and provide the source project for a capable interface. Return the receipt and structured result here. Do not publish, send messages or start ongoing monitoring.
Before you start.
Your interface needs file creation, source bundling and scoped HTTP upload as well as a Recurse connection.
Limits and validation
20 sequential API requests, 300 seconds, 64 KiB input and 256 KiB result per private run.
- A connected Recurse account and sufficient wallet funds are required.
- Authoring requires file creation, source bundling and scoped HTTP upload in addition to MCP.
- Public data may be incomplete, stale, unavailable or sample-limited.
- Recheck current API access, prices and response-field mappings before paid execution.
- A fresh invocation is required for a fresh observation; no background scheduler is included.
This specification was checked against API metadata; a ready-made implementation is not included. Live provider execution has not been tested.