OutSend is a B2B prospecting platform exposed as a remote MCP server. Ask your assistant for "1,000 plumbers in Lyon with verified emails" and it runs the whole job on OutSend: extract local businesses by search query and geographic area with a cost and volume estimate before launching, enrich any finished job with emails, social profiles, reviews, tech stack, and email deliverability checks, chain steps into pipelines (extract, enrich, filter, verify) and schedule recurring monitors that rerun
Scored across 13 tools
Most tools have clearly distinct purposes (scrap vs enrich vs pipeline vs monitor), and descriptions help distinguish them. The 'job' family (create_scrap_job, create_enrichment_job, get_job, get_job_results, list_jobs) clusters somewhat, but status-vs-results and scrap-vs-enrich boundaries are spelled out.
Names follow a clean verb_noun pattern: create_*, get_*, list_*, estimate_scrap. Consistent snake_case throughout with predictable prefixes and no mixed conventions.
13 tools is well within the sweet spot for a scraping/enrichment/pipeline platform. Each tool maps to a distinct operation (create/get/list/estimate) and none feel redundant.
Creation and read coverage is strong (jobs, pipelines, monitors, events, schema), but lifecycle management is missing: no cancel/delete job, no stop/delete pipeline, and no update/delete monitor. Agents can create resources but cannot clean them up, creating dead ends.
Enrich the results of a finished job (async): find emails, scrape reviews, detect social networks, verify emails, tech stack, etc. Returns the new job id.
Inputs: enrichment (required), source_job_id (required), email_mode, max_reviews, +1 more.
Turn a finished job or pipeline into recurring monitoring: outsend re-runs it every N days and computes the diff (new / removed / modified places).
Inputs: name (required), frequency_days (required), source_job_id, source_pipeline_id.
Create AND start a multi-step pipeline (e.g. scrap → emails → verify_emails). Chains are linear: each block feeds the next. Edges use the keys "from" and "to" (not source/target). Valid block types and config fields are described by get_pipeline_schema. Returns pipeline id + initial job ids. Example definition: {"nodes": [{"id": "n1", "type": "scrap", "config": {"queries": ["plombier"], "zones": ["Lyon"], "stop_at": 500}}, {"id": "n2", "type": "emails", "config": {}}], "edges": [{"id": "e1", "from": "n1", "to": "n2"}]}
Inputs: definition (required), name, low_priority.
Launch a Google Maps extraction (async). Returns the job id immediately; the job runs in the background until status=done. stop_at caps the number of unique places. Listings whose phone number carries another country's calling code are dropped by default (exclude_foreign_phones).
Inputs: zones (required), queries (required), country, stop_at, +4 more.
Estimate the cost (EF) and volume of a Google Maps extraction WITHOUT launching it. 1 EF = a full-France scrape for one query type.
Inputs: zones (required), queries (required), country, stop_at, +1 more.
Account event feed (job.completed, pipeline.completed, veille.run_completed…). Cursor-based: since_id takes the last_id of the previous call.
Inputs: limit, types, since_id.
Status, counters and download URL of a job.
Inputs: job_id (required).
Preview of a finished job's rows (max 50 per call, paginate with offset) + total count + CSV download URL for the full export.
Inputs: job_id (required), limit, offset.
Machine contract for building pipelines: valid block types, per-block config schema, and chaining rules accepted by create_pipeline.
Inputs: no named parameters.
Catalog of outsend data modules: what each one needs and produces. Modules listed as coming_soon cannot be launched yet.
Inputs: no named parameters.
The account's recurring monitors.
Inputs: no named parameters.
The account's pipelines with status.
Inputs: no named parameters.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Add one secure layer between your agents and this server.