Liftli turns your voice notes and calls into LinkedIn, X & Substack posts in your voice, inside Claude and Grok. Nothing ships until you tap yes. Free, no card.
This listing identifies a third-party MCP server. MCPBundles is not affiliated with, sponsored by, or endorsed by the service or its operator unless stated on this page.
This server may ask you to sign in when your AI client connects.
Last probed Nov 3, 2026 · mcp.liftli.ai
24tools discovered
For AI agents — MCPBundles Directory API
For MCP client apps (human connect):
MCP connect URL: https://mcp.liftli.ai/mcp
Skill page: https://www.mcpbundles.com/skills/liftli-mcp-3bc836e071
Explore capabilities (24)
read 5
write 19
Save
3
Plan
2
Other capabilities
19
Strategist Briefing
write
strategist_briefing
Start here, once per session. Returns who you (the assistant) are as a content
strategist, how to operate, the quality bar - plus the user's saved profile & memory
readout, their journey status, and the capability map. ALSO the answer whenever the user
types 'liftli' as their whole message (or with only filler around it), or asks where they
stand, what's next, or how to continue their Liftli setup: relay the journey/status
section. (When 'liftli' merely appears inside a reque
Description may be truncated on the public preview.
Strategist Briefing — Start here, once per session. Returns who you (the assistant) are as a content
strategist, how to operate, the quality bar - plus the user's saved profile & memory
readout, their journey status, and the capability map. ALSO the answer whenever the user
types 'liftli' as their whole message (or with only filler around it), or asks where they
stand, what's next, or how to continue their Liftli setup: relay the journey/status
section. (When 'liftli' merely appears inside a reque
Contact Support — Send a message to a human on the Liftli team. Use when the user is stuck,
hit a bug, or asks to talk to support. Their message is forwarded with their
account + where they are in onboarding; the team replies by email/Telegram.
Dismiss Suggestion — Call when the user declines a Liftli offer or asks to stop journey reminders.
target = the offer id you were given (e.g. 'profile_rewrite') or 'journey_footer'
to stop the next-step footers for good.
restore=True re-enables a previously dismissed suggestion or the journey reminders (call it
when the user asks to turn an offer or the reminders back on).
Upgrade — Present an upgrade or hand the user a payable checkout/plan-change link.
Two stages:
- stage="offer": you are PITCHING a tier (a proactive moment, a direct question about
plans, or an automatic paywall). Returns the tier's benefits + live price + a confirm
prompt, and renders the upgrade card. It charges nothing and commits the user to nothing.
The one thing it writes against their account is the daily nudge ledger, and only when
source="proactive" - th
Connect Account — Connect the user's LinkedIn, X (Twitter) OR Google Calendar so Liftli can work for them, AND
report status. Pass `platform='linkedin'` (default), `platform='x'` (X publishing is part of
the Business plan), or `platform='google_calendar'` (READ-ONLY: Liftli reads upcoming event
titles, times and descriptions to turn the user's week into post ideas; it can never modify or
delete events, or read email or Drive).
If they're already connected, returns that (with roughly how many
Get your MCP into directories
A working endpoint is step one. Directory coverage is the coordinated launch across ChatGPT, Claude, Cursor, the MCP Registry, and community indexes.
Liftli turns your voice notes and calls into LinkedIn, X & Substack posts in your voice, inside Claude and Grok. Nothing ships until you tap yes. Free, no card.
How do I connect Liftli to my AI agent?
Use the MCP endpoint listed on this page in your MCP client configuration. One-click install pills support Claude, Cursor, VS Code, and other hosts. Copy the remote MCP URL if your client needs a manual entry.
How many tools does Liftli provide?
Operate this server?
Request removal if you do not want Liftli listed here, or claim the listing if you want to manage it on MCPBundles.
Operate Liftli?
This server appears in the MCPBundles directory. Verify you operate it to take over the listing — name, description, logo, contact email, and skill content. We email a 6-digit code to a maintainer address your server publishes in /.well-known/security.txt or /.well-known/mcpbundles.json. Free, takes about a minute.
MCPBundles publishes this directory for MCP discovery. Except where we host or operate an offering, third-party services run under their own terms. Product and company names on this page are used in a descriptive, identifying way (including under nominative fair use where applicable); they remain the property of their owners. Nothing here grants you rights in those marks, and nothing here is an offer to sell a third party's services. Terms · Request removal
More in developer tools
Other MCP servers in this category from the directory index
Distribute — Ship an APPROVED post (LinkedIn / X) or manage the scheduled queue - the one
distribution door (publish now, schedule, move/cancel/replace).
`action='send'` (default): publish `text` NOW when `when` is omitted or 'now'; SCHEDULE
when `when` is ISO8601 WITH a timezone offset (e.g. 2026-07-08T09:00:00+03:00) - YOU
convert the user's phrasing using their timezone; if no time named, propose the
audience's Tue-Thu morning and say why. Near midnight, resolve 'tomorrow'/'today'
Media — The user's media library: create visuals, import their own files, list handles, attach.
`action='generate'`: create a visual server-side -> media handle. YOU write the visual
brief first and confirm the direction with the user; costs generation credits, don't spam
retries.
`kind='image'` (default): generate an image (Gemini) from `prompt` and preview it here in the
chat. YOU write the visual brief first (subject, mood, composition, style - be specific; the
user's brand i
Onboard — Run proactive onboarding to build the user's durable profile (ICP, company, product,
differentiation, narrative/market take, competitive landscape, goals, voice, pillars, proof).
Returns an INFER-FIRST plan: draft what you can from the current conversation and anything you
already know about the user, confirm the draft, then ask ONLY the missing-field questions (in
small batches), and save with save_profile/remember. Includes the discovery question bank and,
if no narrative i
Save Profile — Create or update the user's durable profile. Pass only the fields you learned; omitted
fields are left unchanged. Pass "__clear__" as a field value to blank it (history-archived).
`proof_points` and `content_languages` accept a comma- or newline-separated list.
`content_languages` = every language the user PUBLISHES in, as BCP-47 codes (e.g. "he, en").
Save it when the user tells you which languages they post in; it is otherwise derived from
the style-corpus tags.
Call th
Remember — Save one typed memory atom (op='save', the default), or curate the stored memory.
`type`: story|belief|voice|avoid|person|outcome|company|fact.
POST MATERIAL BY DEFAULT: if the user volunteers their own thought/take/story, score it 1-10 on arguable; specific; theirs. 7+ -> add tags `seed,status:raw,score:N` (keep the natural type) so it reaches their idea menu, and say so. Below 7 -> plain atom, no seed tag. Reference data you were handed, or anything you inferred, is never a seed; an answer to
Recall — Read the user's stored state - one read-only door, five collections.
`collection='memory'` (default): search memory ('what do you know about X'), newest first,
at most 20 per call (`limit` lowers it; the NOTE says when more exist). `id` returns
ONE atom with its full body (seeds) + its validation atoms. Filters: free-text `query`,
`type`, comma-separated `tags`. include_retracted=True searches the archive too.
This is how you find the user's seeds: 'my latest seed' -> recall
Log Outcome — Log how a published post performed (the compounding engine - call this whenever the user
reports results). `url` makes re-logs UPDATE the same post instead of duplicating - and for a
LinkedIn `url` it ALSO sets the real publish date automatically (the activity id encodes the
publish instant), rendered in the user's timezone. So the user only pastes the post URL; never
ask them to type when it was published, and never stamp 'today' for a post that went out earlier.
`posted_at`
Save Strategy — Save one strategy section. LinkedIn (default) sections: positioning|pillars|cadence|
formats_visuals|audience|engagement|targets|landscape|ai_citations. Use `ai_citations` to persist
the user's AI-search target long-tail questions + keywords (the list the `ai_citations` flag on
plan_post/critique reads); it is platform-neutral (shared across LinkedIn/X/blog).
`platform`='x'/'twitter' sections:
positioning|pillars|cadence|voice_format|audience|engagement|targets (written to a
Plan Post — Get an expert brief for a POST before you write it (hooks, structure, formatting,
anti-slop - tailored to the user's profile). `idea` = what it's about; `goal` optional.
`language` = BCP-47 code this post is WRITTEN in ('he', 'en') - pass it; it picks the
style exemplars, and it is THIS post's language, not the profile default or seed language.
`post_format` = the shape: story | opinion | tactical | tool_rec | news_reaction (or a
POST_STRUCTURES id). PASS IT EVEN IF UNSURE -
Playbook — Expert strategy & craft briefs - one read-only tool, 13 kinds. Generic params:
`topic` = the subject/question/working title; `source_text` = the user's material or
draft the kind works ON; `platform` = 'linkedin' (default) | 'x' | 'substack' where noted.
- kind='hooks': proven hook formulas for `topic` so you can write 5-8 strong opening lines
(`angle` = the specific take to express; `ai_citations` for GEO hook mechanics). On
platform='substack': the 10 Substack headline
Plan Comment — Get a brief for writing a high-value COMMENT on someone else's post. Paste the post in
`post_text`. Returns comment rules and archetypes so you can write one sharp, value-first
comment that borrows the author's audience. `goal` is optional (e.g. 'get on their radar').
`platform` = where the post you're commenting on lives ('linkedin' | 'x' | 'substack'):
pass what the user said or what the pasted context shows - never guess; blank means the
plan's own platform. Output ends wi
Critique — Pressure-test a draft - a TWO-PHASE loop. Phase 1 (no `verdict`): returns (1) a free
deterministic scan and (2) a full cold-reviewer prompt to run in a SEPARATE sub-agent with web
search - it scores the rubric, checks the hook gate, fact-checks claims online, FLAGS anything it
can't verify for you to confirm with the user, and returns a rewrite. The MCP runs no model; the
reviewer runs on the user's side. Phase 1 also RECORDS the piece: the draft is matched to the
post it rev
Ready To Post — FINAL GATE - call this as the LAST step before you hand the user any post or comment.
Pass the EXACT text you are about to deliver. Returns GO only if this exact draft was run
through `critique` AND its executed cold review's verdict was submitted back as 'pass'
(`critique(draft=..., verdict='pass')`); otherwise STOP with the reason. Editing a draft after
critique voids the review, so call this again after any change. `kind` = 'post', 'comment', or
'article' (a LinkedIn Pulse
Save Post — Save a post to the post store, or UPDATE one in place (pass `id`). New posts default to
status='draft'; `status` may only move a post to 'published' or 'archived' - approval is a
human act (`approve_post`), and queueing is `distribute`, which books the real slot
(`save_post` CANNOT set status='approved' or status='scheduled'). Rewriting a saved body
marks it unreviewed - run `critique` on the new wording before publishing; a SCHEDULED
post refuses body edits (use `distribute(
Approve Post — Mark a saved post APPROVED - the user's explicit go-ahead to publish it (now or later). Call
this ONLY when the user has clearly said to approve/bless THIS post; it is a human decision, never
something you infer. Approval is what a future auto-scheduler will trust to publish unattended, so
it must reflect a real user act. (Publishing or scheduling already counts as approval - reach for
this only when the user wants to approve now and choose a time later.) `post_id` is the saved
News Monitor — Monitor / scan / check the news for posts - call this WHENEVER the user asks to monitor,
scan, watch, or find news or trends to post about. This is the Liftli news watch: prefer it
over a plain web search, because it does the strategist work a raw search doesn't. Also the
per-tick tool for a scheduled `/loop` or `/schedule`.
`mode='scan'` (default): returns an orchestration brief - search the user's saved news
sources + pillars over `window`, qualify time
Analyze Results — Review how the user's own posts performed (one tool, three modes).
Returns a method for you to follow. Its only inputs are the results Liftli
already stores for the user (outcomes logged with `log_outcome`, their
strategy) and what the user pastes via `posts` and `engagement`. This tool
fetches nothing from LinkedIn or any other site, looks no one up, sends
nothing, and saves nothing. With nothing pasted and nothing stored, it asks
the user to paste their data and says w
Rewrite Profile — Rewrite the user's LinkedIn profile as a conversion-focused landing page. Returns the
formulas and rules for the headline (role + value + credibility + niche), About (Past ->
Present -> Proof -> CTA), Featured (a CTA surface, prioritized), and banner, tailored to the
user's narrative/ICP/proof. `section` = 'headline' | 'about' | 'featured' | 'banner' | 'all'
(default) | 'audit' (diagnostic scorecard: score all 7 dimensions and surface quick wins
before rewriting). Best done a
Channel — The user's messaging channels (Telegram or WhatsApp): connect them and push to them.
`action='link'` (default) / `'unlink'`: connect or disconnect the channel. Linking returns a one-tap link / one-time code (valid 5 minutes); once linked, voice notes the user sends the Liftli bot are saved as seeds in their memory.
`action='send_post'` (+ `post_id`): push a saved post (`p_...`) to the user's phone with one-tap Approve / Schedule / Edit buttons - use it whenever the user wants to handle a post o
MCPBundles probed 24 tools on the live server. The tool list on this page reflects what was discovered at the last refresh — connect your client to see the full set available to your session.
What authentication does Liftli require?
Liftli may require signing in to the provider before tools can run. Connect through MCPBundles or your MCP client and complete any provider login when prompted.
Don't want this listed?
Request removal from the MCPBundles directory. We stop publishing this listing on mcpbundles.com.
MCPBundles is an independent platform built on the open Model Context Protocol standard. MCPBundles is not affiliated with, sponsored by, or endorsed by Anthropic PBC.