← all tools
Sequencers

LeonarMCP connector for Claude and ChatGPT

Run Leonar (recruiting sequencer: email/WhatsApp sequences, plus the contacts and project pipelines behind them) — create a draft sequence with its steps, edit a draft or paused sequence and its steps, link a sequence to a project, enroll contacts and set their per-contact variables, activate, pause or resume a sequence, read enrollments, conversations and replies, and send a one-off email or WhatsApp message; also list/search/create/update contacts and companies, add notes, tags and tasks, put contacts on a project pipeline and move their stage, and enrich a contact's email or phone. Use when the user mentions Leonar, a Leonar sequence or campaign, or pushing candidates into Leonar. Keys: BYOK-only — each member connects their own Leonar API key.

Leonar is an outreach sequencer, and one of the 55 tools hyreflow connects to Claude, ChatGPT and other AI assistants. Through hyreflow, an AI agent can run 45 actions: 23 that look something up and 22 that create or change records. Leonar runs on your own Leonar account: hyreflow charges no credits for it, and Leonar bills you as before.

Cost per callBYOK · vendor billed
Your own keyrequired
CategorySequencers

How do you connect Leonar to Claude or ChatGPT?

You connect hyreflow once over MCP, and Leonar comes with it, next to the other 54 tools. You do not need to run a separate Leonar MCP server.

  1. 1

    Add hyreflow as a connector in your assistant

    In Claude, ChatGPT or any other MCP client, add a custom connector named hyreflow with the server URL https://recruit.hyreflow.ai/api/v2/mcp. The MCP setup guide on hyreflow.ai has the exact steps for each app.

  2. 2

    Sign in to hyreflow and pick a workspace

    A browser window opens. Sign in, choose the workspace the assistant may act on, and approve. The connection is live.

  3. 3

    Connect your Leonar account

    Leonar runs on your own Leonar account. Connect it on the Integrations page in hyreflow, and hyreflow charges no credits for its calls.

  4. 4

    Ask for the outcome

    Tell Claude or ChatGPT what you want done with Leonar in plain language. The assistant picks the Leonar action that fits and runs it through hyreflow.

Using Cursor, Claude Code or another client? Every one is covered in the MCP setup guide.

What are the alternatives to Leonar?

Other outreach sequencers hyreflow connects: EmailBison, HeyReach, Instantly, Lemlist, Reply.io and SendKit. They run through the same connector, so you can switch, or combine them in one run, without new setup.

Which Leonar actions can an AI agent run through hyreflow?

Through hyreflow, an AI agent can run 45 Leonar actions: 23 that look something up and 22 that create or change records. Leonar runs on your own Leonar account: hyreflow charges no credits for it, and Leonar bills you as before.

Leonar actionWhat it doesCost through hyreflow
Enrich contactLooks something upno hyreflow credits
Search contactsLooks something upno hyreflow credits
Add contact noteCreates or changes a recordno hyreflow credits
Add contact tagCreates or changes a recordno hyreflow credits
Add to pipelineCreates or changes a recordno hyreflow credits
Add to pipeline batchCreates or changes a recordno hyreflow credits
Create companyCreates or changes a recordno hyreflow credits
Create contactCreates or changes a recordno hyreflow credits
Create contacts batchCreates or changes a recordno hyreflow credits
Create sequenceCreates or changes a recordno hyreflow credits
Show all 45 Leonar actions →
Leonar actionWhat it doesCost through hyreflow
Create taskCreates or changes a recordno hyreflow credits
Send messageCreates or changes a recordno hyreflow credits
Activate sequenceCreates or changes a recordno hyreflow credits
Enroll contactsCreates or changes a recordno hyreflow credits
Link project sequenceCreates or changes a recordno hyreflow credits
Pause sequenceCreates or changes a recordno hyreflow credits
Remove contact tagCreates or changes a recordno hyreflow credits
Replace sequence stepsLooks something upno hyreflow credits
Resume sequenceCreates or changes a recordno hyreflow credits
Set project sequenceCreates or changes a recordno hyreflow credits
Unenroll contactsLooks something upno hyreflow credits
Unlink project sequenceCreates or changes a recordno hyreflow credits
Update contactCreates or changes a recordno hyreflow credits
Update enrollment variablesCreates or changes a recordno hyreflow credits
Update pipeline entryCreates or changes a recordno hyreflow credits
Update sequenceCreates or changes a recordno hyreflow credits
List companiesLooks something upno hyreflow credits
List contact notesLooks something upno hyreflow credits
List contactsLooks something upno hyreflow credits
List conversationsLooks something upno hyreflow credits
List enrollmentsLooks something upno hyreflow credits
List messagesLooks something upno hyreflow credits
List pipeline entriesLooks something upno hyreflow credits
List projectsLooks something upno hyreflow credits
List sequence stepsLooks something upno hyreflow credits
List sequencesLooks something upno hyreflow credits
List tagsLooks something upno hyreflow credits
List tasksLooks something upno hyreflow credits
Get companyLooks something upno hyreflow credits
Get contactLooks something upno hyreflow credits
Get enrichment statusLooks something upno hyreflow credits
Get pipeline entryLooks something upno hyreflow credits
Get projectLooks something upno hyreflow credits
Get project sequenceLooks something upno hyreflow credits
Get sequenceLooks something upno hyreflow credits

One credit is priced on the pricing page. The full price list for every tool is in the pricing tables.

Ask for the outcome and your agent composes the run itself. Or start from a skill, a whole pipeline it already knows end to end.

browse all skills →

How does an AI agent know how to use Leonar?

hyreflow ships a written playbook for Leonar, and the agent loads it before the first call. Auth, rate limits, what each call costs, which actions need your approval, and the mistakes worth avoiding. It is the difference between an agent that knows the tool and one that guesses at it.

Read the Leonar docs →
Show the raw playbookwritten for the agent

HOW in lib/leonar.py. Docs: leonar.mintlify.app (published OpenAPI spec).

Leonar is a recruiting sequencer: its sequences are the outreach campaigns (email and WhatsApp steps). It also stores the contacts, companies and project pipelines those sequences run on, so one connection covers both the outreach and the records.

Auth & config (API key)

  • Connect (BYOK): dashboard Integrations → Leonar — paste the API key, then Save and test. CLI: hyreflow byok set leonar --api-key <key>.
  • Get the key in Leonar: in the Leonar V2 app, open Settings → API and create a key. It starts with leo_. Keys from the old V1 app start with pk_live_ and do not work: Hyreflow refuses them before calling Leonar. API access needs an active Leonar subscription (billing_required otherwise), and a feature outside the plan answers plan_upgrade_required.
  • A key belongs to the Leonar user who created it: messages and enrollments go out from that user's connected email/WhatsApp accounts, and a key only reaches what its scopes allow (a read-only key can't write). Each member connects their own key (the owner's key is the workspace key), so what they send goes out from their own accounts and their notes and tasks carry their name; until a member adds theirs, Leonar won't run for them. Revoking the key in Leonar breaks the connection until a new one is pasted.
  • Leonar allows 2000 requests per hour per user, shared by every key that user owns. Hyreflow paces each key at 30 requests/minute, which keeps one key under it; two keys of the same user used at the same time can still hit a 429, which is retried.

Noun mapping

Sequence = an outreach campaign: an ordered list of steps (email, WhatsApp, call, manual and enrichment steps, with delays and branches); an enrollment is one contact in it. Contact = any person (candidate or client contact). Company = account. Project = a recruiting role with its own pipeline of stages, and optionally one default sequence. Pipeline entry = one contact at one stage of one project. Conversation = the message thread with one contact.

Operations

Sequences (campaigns): read list_sequences / get_sequence / list_sequence_steps / list_enrollments, get_project_sequence; build and edit (sends nothing while the sequence is draft or paused) create_sequence, update_sequence, replace_sequence_steps, link_project_sequence / set_project_sequence / unlink_project_sequence, update_enrollment_variables, pause_sequence. Outreach (reaches real people — gate each one): activate_sequence, resume_sequence, enroll_contacts, send_message. Each one is refused unless the payload carries "confirm": true, and none of them runs through batch: send confirm only after the user said yes to that exact call, never to get past the refusal. unenroll_contacts stops a contact's remaining sends. Replies: list_conversations / list_messages. Records: read list_contacts / search_contacts / get_contact, list_contact_notes, list_companies / get_company, list_projects / get_project / list_pipeline_entries / get_pipeline_entry, list_tasks, list_tags, get_enrichment_status; write create_contact, update_contact, add_contact_note, add_contact_tag / remove_contact_tag, create_company, add_to_pipeline, update_pipeline_entry, create_task, enrich_contact. Bulk: create_contacts_batch, add_to_pipeline_batch, plus the generic batch for get/search/update/create calls (see crm-bulk-pushes.md). batch refuses the four sending methods, replace_sequence_steps and enrich_contact.

Sequences: states and what sends

  • A sequence is draft, active, paused or archived. Read its state from status on get_sequence (or list_sequences) before any write below; don't assume it from what you just did.
  • Draft and paused are the non-sending states. create_sequence always creates a draft (no other status is accepted), and a draft sends nothing until activate_sequence. pause_sequence stops an active sequence and its pending tasks until resume_sequence. Leonar only lets you edit a sequence in these two states: update_sequence and replace_sequence_steps answer 409 on an active or archived sequence, so pause it first.
  • Building steps on a draft or paused sequence is not itself the gated act once the user has confirmed the cadence: confirm the state with get_sequence, then write the steps. activate_sequence and resume_sequence are the gate: they start or continue live email/WhatsApp sends from the user's connected accounts to every contact enrolled. Show the sequence, its steps and how many contacts are enrolled, and get a yes before either one. Neither is retried on a server error: re-read the sequence's status before trying again.
  • create_sequence needs name (max 200): {"name": "Staff Frontend — Paris", "project_id": "<project id>", "steps": [{"client_key": "s1", "type": "email", "subject": "Staff Frontend role at Doctolib", "body": "...", "delay_days": 0}, {"client_key": "s2", "type": "whatsapp", "body": "...", "delay_days": 3}]}. At most 50 steps here. project_id links it as that project's sequence (409 if the project already has one). Other fields: priority (normal, high), mode (simple, advanced), owner_ids (1 to 50), schedule_config.
  • replace_sequence_steps(sequence_id, steps) replaces the whole list (at most 100 steps): a step with an id is updated, one without is created, and every existing step you leave out is deleted. Read the current steps with list_sequence_steps, change them, and send the full list back. Leonar refuses it (409) if a step you remove still has pending tasks. On a paused sequence, the new steps are what enrolled contacts get when it resumes.
  • update_sequence changes only name, priority, mode, owner_ids and schedule_config. The state changes only through activate_sequence, pause_sequence and resume_sequence.
  • A project's default sequence: get_project_sequence reads it. link_project_sequence(project_id, sequence_id) links an existing one, or link_project_sequence(project_id, sequence={...}) creates a new draft (same fields as create_sequence) and links it; both answer 409 if the project already has one. set_project_sequence replaces the link, and unlink_project_sequence only detaches it: the sequence is kept.
  • Per-contact variables after enrolling: update_enrollment_variables(sequence_id, contact_id, {"custom_variable_2": "..."}) changes one enrolled contact's values (keys you leave out are kept, null clears one). It never sends anything itself; steps not yet processing use the new values, and a step already processing or a manual task already created keeps the old ones.
  • enroll_contacts starts real outreach as soon as the sequence is active. Pass either contact_ids or contacts ([{"contact_id": ..., "custom_variables": {...}}]), not both, at most 500 per call either way (unenroll_contacts too). Leonar skips contacts marked do-not-contact, already in this sequence or active in another one, and returns the counts (enrolled, skipped_blacklisted, skipped_already_enrolled, skipped_active_elsewhere, skipped_not_found): report them. Enrolling always stays gated, whatever state the sequence is in: confirm the sequence and the list with the user first.

Notes & guardrails

  • Paging: every list_* takes limit (1 to 100, default 50) and offset (from 0). A larger limit is rejected, not trimmed. Keep raising offset by limit while the response's meta.has_more is true; meta.total is the full count. search_contacts takes limit and offset in its request body: limit must be at least 1, and Leonar documents no maximum for it, so page with meta.has_more rather than assuming a page size.
  • Find before you create: search_contacts takes query (full text) or name, title, company, location, tags, project_id, sequence_id. Search first and update the match with update_contact instead of creating a second record.
  • create_contact needs first_name: {"first_name": "Sophie", "last_name": "Martin", "title": "Senior Software Engineer", "current_company": "Doctolib", "emails": [{"email": "[email protected]", "type": "work"}], "linkedin_profile": "https://www.linkedin.com/in/sophie-martin", "tags": ["frontend"]}. Up to 5 emails and 5 phones. tags are names and are created if they don't exist. custom_data is keyed by the workspace's custom field keys.
  • update_contact changes only the fields you send. In custom_data, a key set to null is removed and keys you leave out are kept.
  • Tags on an existing contact go by id: read them with list_tags, then add_contact_tag(contact_id, tag_id). remove_contact_tag only detaches the tag.
  • Pipeline stages are ids, not names. Read a project's stages with get_project (data.pipeline.stages), then pass the stage's id as stage_id. add_to_pipeline(project_id, contact_id) without a stage_id lands the contact in the first stage. Move a contact with update_pipeline_entry(entry_id, {"stage_id": "<stage id>"}); only the fields you send change.
  • Custom variables (on add_to_pipeline, update_pipeline_entry, enroll_contacts and update_enrollment_variables) are the fixed keys custom_variable_1 to custom_variable_5, each at most 2000 characters. Any other key is rejected. They fill the {{custom_variable_N}} placeholders in the sequence's messages.
  • Enrichment spends the customer's Leonar enrichment credits: enrich_contact(contact_id, type) with type = work_email, personal_email or phone returns a request_id; poll get_enrichment_status(request_id) until status is completed or failed. Confirm before enriching a list, one contact per call: batch refuses enrich_contact.
  • send_message sends immediately to a real person from the user's connected account: channel is email (needs subject) or whatsapp. Show the exact message and get a yes before every send. It is not retried on a server error, because the message may already have gone out: check list_conversations for the contact before sending again.
  • No record delete is exposed. The one write that removes anything is replace_sequence_steps, which deletes the steps you leave out; the generic batch refuses it.
  • Production writes: pilot one record, confirm, then bulk.

Handoff

Outreach: enriched candidates → create_contact (or create_contacts_batch) → add_to_pipeline on the role's project (or add_to_pipeline_batch) → the project's sequence (get_project_sequence, or create_sequence as a draft with its steps) → after the user approves, enroll_contacts and, for a draft, activate_sequence. Replies come back as list_conversations / list_messages and as enrollment status (replied, bounced) on list_enrollments; log call notes with add_contact_note and follow-ups with create_task.

Bulk writes & rate limits

For large pushes use create_contacts_batch and add_to_pipeline_batch (each payload {contact_id, stage_id?, custom_variables?}, plus project_id once for the whole list). They pace calls to the 30 req/min default and return {created, failed, total, elapsed_ms}. At that pace 500 contacts take about 17 minutes. See crm-bulk-pushes.md for the cross-CRM guide.

Callable surface

Call via the CLI: hyreflow tools execute leonar <method> --payload '{...}' (preview with --dry-run; hyreflow tools get leonar <method> returns the live contract + cost). Base: https://app.leonar.app/api/v1. Only the methods listed below are callable.

  • activate_sequence(sequence_id: str, *, confirm: bool = False) -> Any — POST /sequences/{id}/activate — switch a draft sequence on: it starts sending its email/WhatsApp steps to every enrolled contact. 409 if it is not a draft. Refused without confirm=true: pass it only after the user approved the sequence, its steps and who is enrolled.
  • add_contact_note(contact_id: str, content: str, *, is_pinned: bool | None = None, project_id: str | None = None) -> Any — POST /contacts/{id}/notes — add a note (1-10000 characters) to a contact, optionally scoped to a project.
  • add_contact_tag(contact_id: str, tag_id: str) -> Any — POST /contacts/{id}/tags — attach an existing tag (by id, see list_tags) to a contact.
  • add_to_pipeline(project_id: str, contact_id: str, *, stage_id: str | None = None, custom_variables: dict | None = None) -> Any — POST /projects/{id}/entries — add an existing contact to a project's pipeline, at stage_id or else the first stage. custom_variables = per-contact custom_variable_1..5 for the project's sequence.
  • add_to_pipeline_batch(payloads: list[dict], *, project_id: str | None = None, stop_on_error: bool = False) -> dict — POST /projects/{id}/entries for many contacts, rate-limited. Each payload is {contact_id, stage_id?, custom_variables?} plus project_id, or pass project_id once for every payload that lacks one.
  • batch(method: str, calls: list, *, stop_on_error: bool = False) -> dict — Batched get/search/update/create: run one method over many inputs, rate-limited. Each call is a value -> method(value), or {"args": [...], "kwargs": {...}} for multi-arg methods like update_contact(id, payload). Destructive methods (delete) are refused, and so are replace_sequence_steps, which deletes the steps it leaves out, the sending methods (enroll_contacts, activate_sequence, resume_sequence, send_message), which run one confirmed call at a time, and enrich_contact, which spends Leonar enrichment credits per row. See lib/_base.run_ops.
  • create_company(payload: dict) -> Any — POST /companies — create a company. name is required; also domain, industry, company_size, description, location, logo_url, phone, linkedin_url, annual_revenue, owner_id, status_id, custom_data.
    • requires: name
    • example: {"name":"Doctolib","domain":"doctolib.com","industry":"Health Tech","company_size":"1001-5000","location":"Paris, France"}
  • create_contact(payload: dict) -> Any — POST /contacts — create a contact. first_name is required; also last_name, email or emails[] ({email, type}, max 5), phone or phones[] (max 5), title, current_company, location, linkedin_profile, contact_type[], tags[] (names, created if missing), source, company_id, custom_data, global_status, do_not_contact, years_experience, skills[], languages[].
    • requires: first_name
    • example: {"first_name":"Sophie","last_name":"Martin","title":"Senior Software Engineer","current_company":"Doctolib","location":"Paris, France","emails":[{"email":"[email protected]","type":"work"}],"linkedin_profile":"https://www.linkedin.com/in/sophie-martin","tags":["frontend"]}
  • create_contacts_batch(payloads: list[dict], *, stop_on_error: bool = False) -> dict — POST /contacts for many records, one at a time, respecting the Leonar rate limit.
  • create_sequence(payload: dict) -> Any — POST /sequences — create a sequence as a draft (it sends nothing until activate_sequence). name is required (max 200); also priority (normal, high), mode (simple, advanced), project_id (link it as that project's sequence; 409 if the project already has one), owner_ids (1-50), schedule_config, and steps (max 50): [{client_key, type (email, whatsapp, call, email_enrichment, phone_enrichment, manual_step), name, subject, body, content_template, delay_days (0-365), delay_hours (0-23), execution_mode (automatic, manual), sender_ids, parent_step_id (a client_key), branch_condition, entry_conditions}].
    • requires: name
    • example: {"name":"Staff Frontend \u2014 Paris","project_id":"a1b2c3d4-5678-90ab-cdef-1234567890ab","steps":[{"client_key":"s1","type":"email","subject":"Staff Frontend role at Doctolib","body":"...","delay_days":0},{"client_key":"s2","type":"whatsapp","body":"...","delay_days":3}]}
  • create_task(payload: dict) -> Any — POST /tasks — create a task. title is required (max 255); also description, type (call, email, whatsapp, todo, reminder, meeting, interview), priority (low, medium, high), due_date (YYYY-MM-DD), assigned_to, contact_id, company_id, deal_id, project_id.
    • requires: title
    • example: {"title":"Call Sophie about the Doctolib role","type":"call","priority":"high","due_date":"2026-11-10","contact_id":"a1b2c3d4-5678-90ab-cdef-1234567890ab"}
  • enrich_contact(contact_id: str, type: str) -> Any — POST /contacts/{id}/enrich — start finding a contact's work_email, personal_email or phone. Runs async on the customer's Leonar enrichment credits; returns request_id to poll with get_enrichment_status.
  • enroll_contacts(sequence_id: str, contact_ids: list[str] | None = None, *, contacts: list[dict] | None = None, confirm: bool = False) -> Any — POST /sequences/{id}/enroll — start outreach to contacts. Pass exactly one of contact_ids (max 500) or contacts ([{contact_id, custom_variables: {custom_variable_1..5}}]). Leonar skips blacklisted, already-enrolled and active-elsewhere contacts and returns the counts. Refused without confirm=true: pass it only after the user approved this sequence and this list.
  • get_company(company_id: str) -> Any — GET /companies/{id} — one company with its contact and deal counts.
  • get_contact(contact_id: str) -> Any — GET /contacts/{id} — one contact with its positions, pipeline entries and active enrollments.
  • get_enrichment_status(request_id: str) -> Any — GET /enrichment/{requestId} — poll an enrichment: status pending, processing, completed or failed, plus result.
  • get_pipeline_entry(entry_id: str) -> Any — GET /pipeline-entries/{id} — one pipeline entry.
  • get_project(project_id: str) -> Any — GET /projects/{id} — one project with its pipeline stages (data.pipeline.stages, ids for stage_id) and entry counts.
  • get_project_sequence(project_id: str) -> Any — GET /projects/{id}/sequences — the project's default sequence link (404 if the project has none).
  • get_sequence(sequence_id: str) -> Any — GET /sequences/{id} — one sequence with its steps and enrollment counts.
  • link_project_sequence(project_id: str, sequence_id: str | None = None, *, sequence: dict | None = None) -> Any — POST /projects/{id}/sequences — give a project its default sequence: pass exactly one of sequence_id (link an existing sequence) or sequence (create a new draft, same fields as create_sequence, and link it). 409 if the project already has one; use set_project_sequence to replace it.
  • list_companies(*, limit: int = 50, offset: int = 0, **filters) -> Any — GET /companies — one page of companies (limit <= 100). Filters: search (name or domain), industry, owner_id, status_id, updated_after, updated_before, sort (created_at, updated_at, name).
  • list_contact_notes(contact_id: str, *, limit: int = 50, offset: int = 0, project_id: str | None = None) -> Any — GET /contacts/{id}/notes — a contact's global notes, or only those on one project with project_id.
  • list_contacts(*, limit: int = 50, offset: int = 0, **filters) -> Any — GET /contacts — one page of contacts (limit <= 100), archived ones excluded by default. Filters: search (name, company or title), contact_type, company_id, location, global_status (new, contacted, replied, bounced, do_not_contact, archived, all), created_after, created_before, updated_after, sort (created_at, updated_at, first_name, last_name, last_contacted_at; - prefix = descending).
  • list_conversations(*, limit: int = 50, offset: int = 0, **filters) -> Any — GET /conversations — one page of email/WhatsApp conversations. Filters: channel (email, whatsapp), status, unread_only, contact_id, sort (last_message_at, created_at, unread_count).
  • list_enrollments(sequence_id: str, *, limit: int = 50, offset: int = 0, **filters) -> Any — GET /sequences/{id}/enrollments — the contacts enrolled in a sequence. Filters: status (active, paused, finished, replied, bounced, all), search (first or last name), contact_id.
  • list_messages(conversation_id: str, *, limit: int = 50, offset: int = 0) -> Any — GET /conversations/{id}/messages — the messages in one conversation.
  • list_pipeline_entries(project_id: str, *, limit: int = 50, offset: int = 0, stage_id: str | None = None, sort: str | None = None) -> Any — GET /projects/{id}/entries — the candidates in a project's pipeline, optionally one stage (stage_id is the stage UUID, not its name). sort: added_at or moved_at, - prefix = descending.
  • list_projects(*, limit: int = 50, offset: int = 0, **filters) -> Any — GET /projects — one page of recruiting projects (limit <= 100). Filters: status (active, archived, all; default active), owner_id, job_status (draft, published, closed), job_enabled, job_visible_on_client_site, search, updated_after, updated_before, sort (created_at, updated_at, name).
  • list_sequence_steps(sequence_id: str) -> Any — GET /sequences/{id}/steps — the full editable steps: content templates, sender config, branching and delays (the shape replace_sequence_steps takes back).
  • list_sequences(*, limit: int = 50, offset: int = 0, **filters) -> Any — GET /sequences — one page of outreach sequences (limit <= 100). Filters: status, search, updated_after, updated_before, sort (created_at, updated_at, name, status).
  • list_tags(*, scope: str | None = None) -> Any — GET /tags — the workspace's tags (ids for add_contact_tag); scope = global, contact, company, deal or project.
  • list_tasks(*, limit: int = 50, offset: int = 0, **filters) -> Any — GET /tasks — one page of tasks. Filters: status (todo, done, skipped, all), type (call, email, whatsapp, todo, reminder, meeting, interview), assigned_to, contact_id, company_id, deal_id, updated_after, updated_before, sort (created_at, updated_at, due_date, priority, status).
  • pause_sequence(sequence_id: str) -> Any — POST /sequences/{id}/pause — pause a sequence and its pending tasks; nothing more is sent until resume_sequence. 409 if it cannot be paused from its current state.
  • remove_contact_tag(contact_id: str, tag_id: str) -> Any — DELETE /contacts/{id}/tags — detach a tag from a contact (tag_id query param); the tag itself stays.
  • replace_sequence_steps(sequence_id: str, steps: list[dict]) -> Any — PUT /sequences/{id}/steps — replace the whole step list of a draft or paused sequence (1 to 100 steps; an empty list is refused).
  • resume_sequence(sequence_id: str, *, confirm: bool = False) -> Any — POST /sequences/{id}/resume — resume a paused sequence and its pending tasks: sending to its enrolled contacts continues. 409 if it is not paused. Refused without confirm=true: pass it only after the user approved resuming.
  • search_contacts(*, limit: int = 50, offset: int = 0, **filters) -> Any — POST /contacts/search — full-text search with combined filters (a read). Body fields: query, name, title, company, location, contact_types[], tags[], created_after, created_before, last_contacted_after, last_contacted_before, salary_min, salary_max, project_id, sequence_id, sort_by, sort_direction (asc|desc).
  • send_message(contact_id: str, channel: str, content: str, *, subject: str | None = None, sender_account_id: str | None = None, recipient_phone: str | None = None, confirm: bool = False) -> Any — POST /messages — send an email or WhatsApp message to a contact now, from the user's connected account (sender_account_id, auto-selected if omitted). Email needs subject. Creates the conversation if there is none. Refused without confirm=true: pass it only after the user approved this message.
  • set_project_sequence(project_id: str, sequence_id: str) -> Any — PUT /projects/{id}/sequences — set or replace the project's default sequence with an existing one.
  • unenroll_contacts(sequence_id: str, contact_ids: list[str]) -> Any — POST /sequences/{id}/unenroll — remove contacts from a sequence (at most 500 ids per call); returns unenrolled and not_found counts.
  • unlink_project_sequence(project_id: str) -> Any — DELETE /projects/{id}/sequences — detach the project's default sequence; the sequence itself is not deleted.
  • update_contact(contact_id: str, payload: dict) -> Any — PUT /contacts/{id} — update a contact; only the fields sent change. custom_data merges: sent keys are upserted, keys set to null are removed, omitted keys are kept.
  • update_enrollment_variables(sequence_id: str, contact_id: str, custom_variables: dict) -> Any — PATCH /sequences/{id}/enrollments — change one enrolled contact's custom_variable_1..5 (found by contact_id); omitted keys are kept, null clears one. Never sends: steps not yet processing read the new values, a step already processing or a manual task already created keeps the old ones.
  • update_pipeline_entry(entry_id: str, payload: dict) -> Any — PATCH /pipeline-entries/{id} — partial update; omitted fields are kept. Body: stage_id (move the contact to that stage), rejection_reason_id (nullable), custom_variables (custom_variable_1..5, nullable).
  • update_sequence(sequence_id: str, payload: dict) -> Any — PATCH /sequences/{id} — change a draft or paused sequence's name, priority, mode, owner_ids or schedule_config (only those fields). An active or archived sequence is refused (409): pause it first.

Questions about Leonar and hyreflow

Can I use Leonar inside Claude or ChatGPT?

Yes. hyreflow is a connector for Claude, ChatGPT and other AI assistants. Once hyreflow is connected, you ask for the outcome in plain language and the assistant calls Leonar for you. hyreflow exposes 45 Leonar actions this way.

Do I need my own Leonar account to use it with hyreflow?

Yes. Leonar runs on your own Leonar account. You connect it to hyreflow once, hyreflow charges no credits for the calls, and Leonar bills you as before.

What does Leonar cost through hyreflow?

Leonar runs on your own Leonar account: hyreflow charges no credits for it, and Leonar bills you as before.

How do I connect Leonar to an AI agent over MCP?

Add the hyreflow MCP server to your assistant and sign in. hyreflow calls the Leonar API directly and exposes its actions as tools the assistant can run, so there is no separate Leonar server to install.

Connect hyreflow to your assistant →

Put Leonar to work from one prompt

Connect hyreflow to Claude or ChatGPT and ask for the outcome. It reads the playbook above before it makes a single call.