# Autobound MCP connector

The Autobound connector brings companies, contacts, buyer signals, and personalized outreach content into Claude, ChatGPT, Cursor, and the Elaichi Agent, so each person can research an account and draft a tailored message inside their own Autobound access.

Source: https://elaichi.ai/connectors/autobound/

## Facts

| | |
| --- | --- |
| Application | Autobound |
| Category | Sales Enablement |
| AI tools | 35 |
| Authentication | Connects with an API key |
| Needs your own OAuth app | No |
| MCP endpoint | https://api.elaichi.ai/mcp |
| Works with | Claude, ChatGPT, Cursor, any MCP client, and the Elaichi Agent |
| Tools advertised by name | No. 35 tools is past the 30-tool threshold, so clients use `search_tools` and `execute_tool` |

## What you can ask once Autobound is connected

- Which Autobound signals hit my accounts this week?
- Summarize the timeline for this Autobound company.
- List Autobound contacts added at accounts with new signals.

## Connect Autobound in Elaichi

This happens once for the organization, before any client is involved.

1. Open Connections, choose Add connection, and pick Autobound.
2. Optionally set Share with, then press Connect.
3. Paste an Autobound API key. One person generates a token in Autobound and pastes it once. Everyone else works through Share with, and never sees it.

Credentials are vaulted and nobody, including the AI, reads them back. The connection becomes a toolbox immediately, so you can curate which Autobound tools are exposed, rename them, or freeze arguments before anyone points a client at it.

## Connect Autobound to Claude

Endpoint: https://api.elaichi.ai/mcp

1. Open Customize, then Connectors.
2. Press Add.
3. Name it, paste the MCP server URL, then Continue.
4. Sign in and approve.

On Team and Enterprise, an Owner adds it once. Everyone else turns it on for themselves.

## Connect Autobound to ChatGPT

Endpoint: https://api.elaichi.ai/mcp

1. Open Plugins, then press the + button.
2. Name it and paste the endpoint into Server URL.
3. Leave Authentication on OAuth, then tick the risk acknowledgement.
4. Press Create, then sign in and approve.

Works on the web today. The plugin directory lives at chatgpt.com/plugins.

## Connect Autobound to Cursor

Endpoint: https://api.elaichi.ai/mcp

1. Open `~/.cursor/mcp.json`.
2. Add the endpoint under `mcpServers`.
3. Reload Cursor, then sign in and approve.

Set up per machine, so repeat it on each computer you work from.

## Connect Autobound to any MCP client

Endpoint: https://api.elaichi.ai/mcp

1. Add the endpoint as a remote MCP server.
2. Sign in and approve.

The Elaichi Agent already has these tools, with nothing to set up.

## What the consent screen decides

Only Read is granted by default, which is not enough to call a Autobound tool. Over MCP there is no trusted place to confirm a write in the moment, so the consent screen is the standing approval rather than a formality. Grant Read and Run tools. Think hard before granting Delete, which reaches into connected apps and cannot be undone.

## What teams do with Autobound through Elaichi

### Draft a personalized email before the call

Sales. Ask for personalized content for a contact and get an opener built from that person's recent signals, ready to paste into your sequence.

### Find accounts that just showed intent

Sales Development. Search Autobound signals for funding rounds, hiring spikes, or leadership changes across your target list, then pull the matching companies and contacts in one go.

### Load a campaign list into Autobound

Marketing. Hand over a list of companies and contacts from an event or webinar and have them created in bulk so the team can start personalizing straight away.

### Catch up on a customer's timeline

Account Management. Pull a company or contact timeline to see what has happened at the account since your last touchpoint before a renewal conversation.

### Keep an eye on credit spend

Revenue Operations. Check remaining credits and what each type of request costs so nobody is surprised at the end of the month.

### Export buyer intent for the pipeline review

Sales Leadership. Create a buyer intent export across your accounts and summarize which segments are heating up before the weekly pipeline meeting.

## Frequently asked questions

### How do I connect Autobound to Claude?

Connect Autobound in Elaichi first: pick Autobound from the catalog and paste in an API key from your Autobound settings, since that is how Autobound signs in. There is no OAuth application to register and no client ID or secret to generate. Then open Claude, go to Customize, then Connectors, then Add, and paste https://api.elaichi.ai/mcp as the endpoint. Claude will ask you to sign in to Elaichi as yourself, and Autobound is ready to use.

### Does Autobound work with ChatGPT and Cursor as well as Claude?

Yes. Once Autobound is connected in Elaichi, the same endpoint, https://api.elaichi.ai/mcp, works in Claude, ChatGPT, Cursor, any other MCP client, and the Elaichi Agent. You connect Autobound once and every client you use picks it up.

### What can an AI agent actually do with my Autobound data?

With Autobound connected, an agent can look up a company or contact, read its timeline, search signals like funding, hiring, or new executives, and draft personalized outreach content for a specific person. It can also add companies and contacts in bulk, create a buyer intent export, and tell you how many credits you have left and what each request costs. Because Autobound has a large set of actions, short concrete asks such as "find signals for Acme this month" work better than long instructions.

### Does connecting Autobound give the AI everything in my workspace?

No. Every request to Autobound runs as the person who signed in, so the AI sees only the companies, contacts, signals, and credits that person can already see in Autobound. Elaichi can narrow that further with roles and restrictions, but it can never grant more than the person's own Autobound access.

### Can my team share one Autobound connection?

Yes. One person connects Autobound with an API key in Elaichi and shares the connection with a team, and nobody else ever handles the key. Each teammate still signs in to Elaichi as themselves, so every draft, search, or export in Autobound is logged under the name of the person who did it.

### Can I stop an agent from deleting or changing things in Autobound?

Yes. In Elaichi you can restrict Autobound action by action, for example allowing searches and drafts while blocking creating companies, deleting webhooks, or spending credits on exports. A blocked action is never shown to Claude, ChatGPT, Cursor, or any other client, so no prompt can reach it.

### What happens to an Autobound connection when someone leaves?

When you offboard a person in Elaichi, their access to Autobound through every client ends at once. If they had connected a shared Autobound account, it keeps working for everyone else on the team. Disconnecting Autobound once in Elaichi removes it from Claude, ChatGPT, Cursor, and every other client at the same time.

## All 35 Autobound tools

Every tool below is callable through https://api.elaichi.ai/mcp once Autobound is connected, subject to the toolbox it is in and the restrictions on the caller.

- **Get single Autobound account by ID** (Get). Get the authenticated Autobound account's details. Returns: id and account-specific attributes; the full field breakdown depends on the account's configuration — consult the Autobound API docs for the field-level schema.
- **Get single Autobound credit by ID** (Get). Get the current running credit balance for the Autobound workspace. Returns: balance, used, granted, and since. This endpoint is free and consumes no credits.
- **Get single Autobound company by ID** (Get). Enrich a single company in Autobound with signal data via POST /v1/companies/enrich. Returns enriched company signal data; the specific request body and response fields are not enumerated in the available source documentation for this endpoint — consult the Autobound API docs for the full schema.
- **Autobound companies bulk create** (Create). Bulk-enrich companies in Autobound with signal data via POST /v1/companies/bulk. Returns bulk enrichment results; the specific request body and response fields are not enumerated in the available source documentation for this endpoint — consult the Autobound API docs for the full schema.
- **Create a Autobound company** (Create). Enrich a company in Autobound with signal data via POST /v1/companies/enrich. Returns enriched company signal data; the specific request body and response fields are not enumerated in the available source documentation for this endpoint — consult the Autobound API docs for the full schema.
- **List all Autobound companies** (List). Search Autobound companies by signal type to discover distinct companies that have one or more matching signals, sorted by most-recent signal. Returns: domain, name, linkedin_url, signal_count, most_recent_signal_at, signals. Required: signal_types. Entity filters (industries, employee_count_min/max, seniority, department, job_title) cap the detectable window to the most recent 14 days. Costs 2 credits per company returned.
- **Get single Autobound company timeline by ID** (Get). Get the Twitter/X timeline data for a company in Autobound. Returns engagement metrics, account metadata, and AI-enriched fields including num_likes, num_reposts, company_twitter_handle, company_twitter_followers, summary, tags, pain_points, and initiatives. Required: id.
- **Create a Autobound contact** (Create). Enrich a contact in Autobound with B2B buying signals by email or LinkedIn URL. Returns `contact`, `company`, `contact_signals`, `signal_summary`, `total`, and `coverage`. Provide at least one of `contact_email` or `contact_linkedin_url`. Costs 2 credits per signal returned.
- **Autobound contacts bulk create** (Create). Look up shared experiences between sellers and a contact in Autobound. Returns `contactEmail`, `contactLinkedinUrl`, `companyUrl`, `companyLinkedinUrl`, and a `data` array of shared experience objects covering company overlap, city overlap, and school overlap.
- **List all Autobound contacts** (List). Search Autobound contacts that have matching B2B buying signals, sorted by most-recent signal. Returns one record per contact including `email`, `name`, `title`, `company`, `signal_count`, `most_recent_signal_at`, and `signals`. Costs 2 credits per contact returned. Applying entity filters (`industries`, `seniority`, `department`, `job_title`, `employee_count_min`, `employee_count_max`) caps the search window to the most recent 14 days.
- **List all Autobound contact timelines** (List). List contact timeline events in Autobound, returning shared experiences between your team members and a contact. Returns: contactEmail, contactLinkedinUrl, companyUrl, companyLinkedinUrl, and a data array of shared-experience objects each containing subtype, userLinkedinUrl, contactLinkedinUrl, userName, contactName, and experience-specific fields.
- **Get single Autobound signal by ID** (Get). Get a single Autobound signal by its id. Returns: signal_id, signal_type, signal_subtype, signal_name, detected_at, association, company, contact, and data. Required: id. This endpoint has no credit cost.
- **List all Autobound signals** (List). List trending Autobound signals from the real-time B2B signal feed. Returns: signal_id, signal_type, signal_subtype, signal_name, detected_at, association, company, contact, and data per signal.
- **Autobound signals search** (Search). Search Autobound signals using structured filters. Returns matching signal records including signal_id, signal_type, signal_subtype, signal_name, detected_at, association, company, contact, and data.
- **List all Autobound signal types** (List). List all signal types in Autobound — the canonical catalog of type slugs accepted as filter values by other endpoints. Returns: type, association, description, count, refresh_cadence, refresh_hours. Pass include_counts=true to receive per-type row counts and metadata alongside each slug.
- **Get single Autobound signal type by ID** (Get). Get detection-date histogram data for a single Autobound signal type. Returns: type, interval, history (array of period/count buckets), total. Required: id. Accepts both canonical slugs (e.g. sec-8k) and legacy aliases (e.g. 8k).
- **List all Autobound webhooks** (List). List Autobound webhooks configured for real-time push delivery of intent signals. Returns: id.
- **Create a Autobound webhook** (Create). Create an Autobound webhook to enable real-time push delivery of intent signals to your endpoint. Returns: id.
- **Update a Autobound webhook by ID** (Update). Update an Autobound webhook configuration by id. Returns: id. Required: id.
- **Delete a Autobound webhook by ID** (Delete). Delete an Autobound webhook by id. Returns an empty 204 response on success. Required: id.
- **Create a Autobound buyer intent export** (Create). Submit an async bulk buyer-intent export job in Autobound. Returns: export_id and status. Required: topic_ids. Supports up to 1,000 topic IDs per job; costs 10 credits per topic. Use the returned export_id to poll job status via the get method.
- **Get single Autobound buyer intent export by ID** (Get). Poll the status of a buyer-intent export job in Autobound by id. Returns: export_id, status, and download_url (populated when the job is ready for download). Required: id.
- **Create a Autobound personalized content** (Create). Generate hyper-personalized outreach content in Autobound across formats such as email, sequence, call script, connection request, SMS, opener, or custom. Returns generated content items including subject, content body, modelUsed, and insightsUsed. Required: contentType, plus either contactEmail or contactLinkedinUrl, and either userEmail or userLinkedinUrl. When contentType is 'sequence', each email consumes 2 credits; n is not compatible with 'sequence'.
- **List all Autobound credit costs** (List). List the per-action credit cost table for your Autobound workspace. Returns: action, cost, description, chargeMode for each pricing row, plus optional floor and base fields when applicable to your workspace's pricing schedule.
- **List all Autobound logs** (List). List recent API request logs in Autobound covering the last 30 days (free, no credits consumed). Returns: type, method, path, status, latency_ms, timestamp for each log entry. Filter by endpoint path, HTTP method, status code range, or timestamp window using since/before.
- **Get single Autobound stat by ID** (Get). Get Autobound corpus statistics including aggregate signal counts broken down by type. Returns: total_signals, total_types, signal_types (array of per-type slug/count pairs), and last_detected_at. No required parameters; this endpoint is free and consumes no credits.
- **Get single Autobound health by ID** (Get). Get the overall service health of Autobound. Returns: status, timestamp, version, services, and uptime. No authentication required.
- **List all Autobound buyer intent topics** (List). List Autobound buyer-intent topics from a catalog of ~38K entries. Supports fuzzy search via q and filtering by category, sub_category, or is_product. Returns: topic_id, category, subcategory, topic, description, is_product. No credit cost.
- **Get single Autobound buyer intent filter by ID** (Get). Get the valid filter enum values for buyer-intent contact searches in Autobound. Returns: seniority, department, revenue_buckets, size_buckets, country, and categories (intent topic categories with subcategories and topic counts). No required parameters; this endpoint is free and does not consume credits.
- **List all Autobound buyer intent companies** (List). List Autobound companies surging on specified buyer-intent topics, drawn from the weekly pre-scored intent snapshot and ranked by composite intent score. Returns: topic_id, intent_score, stage_label, signal_week, topic (name, category, subcategory, is_product), and company. Required: topic_ids (1–10 topic IDs). Credit cost: 1 per company-topic row returned.
- **Autobound buyer intent companies bulk create** (Create). Submit an asynchronous bulk export of company intent rows for up to 1,000 topics in Autobound. The job is queued immediately and streams results as JSONL to cloud storage. Returns: export_id (cexp_-prefixed), status, credits_charged, topics, credits_per_topic, base, and created_at. Poll GET /v1/buyer-intent/exports/{exportId} with the returned export_id for download. Required: topic_ids.
- **List all Autobound buyer intent contacts** (List). List Autobound contacts showing buyer intent for one or more topics. Returns each contact's topic_id, intent_score, contact details (email, job_title, linkedin_url, opted_out), company firmographics (name, domain, industry, revenue), and filters (seniority, department). Required: topic_ids. Costs 1 credit per contact returned; max 250 per page.
- **Create a Autobound buyer intent contact export** (Create). Submit an asynchronous bulk contact intent export job in Autobound. Returns immediately with HTTP 202 including export_id, status, credits_charged, topics, credits_per_topic, base, and created_at; poll GET /v1/buyer-intent/exports/{exportId} until status is `ready` (24-hour signed download URL) or `failed`. Required: topic_ids. Supports up to 1,000 topic IDs and up to 10,000,000 exported rows.
- **Get single Autobound company intent timeline by ID** (Get). Get the intent-score timeline for a single company in Autobound across all buyer-intent topics, drawn from the weekly snapshot. Returns: company (domain, name, industry, employee_count, linkedin_url), timeline (array of per-topic entries with topic_id, composite_score, stage_label, signal_week), and count. An unknown domain returns 200 with an empty timeline and null firmographics. Required: id (company domain, max 253 chars). Credit cost: free.
- **Autobound monitoring pool records bulk create** (Create). Add 1–100 companies or contacts to the Autobound signal monitoring pool in a single batch. Returns per-record results including index, status, and id (pool entry UUID for accepted records). Required: records. All-or-nothing validation — if any record fails, the entire batch is rejected and nothing is added. Costs 10 credits per net-new record; duplicates are free.