Live job intelligence
Founding offer — lock $8/mo for life. First 20 subscribers only.

The best AI-agent gigs vanish in hours. We score every one.

AgentScout tails the freelance marketplaces for AI-agent and automation work, scores every gig 1 to 10 with Claude, and alerts you to the two or three a day actually worth your time — in your dashboard, on Telegram, or in Slack. Found one yourself? One click scores it before you spend a connect.

See how it works
free private beta · full scored alerts · no card, no code · prefer Telegram?
agentscout · sample feed demo
RAG assistant over 10k support docs
Freelancer · $2,000 fixed
9/10
Voice-agent integration build (n8n + Twilio)
n8n community · $20–50/hr
8/10
Contract: agentic AI engineer (6 months)
LinkedIn · contract
7/10
Ship an agent MVP for a seed startup
Reddit · equity / rev-share
5/10
Illustrative — 128 scanned · 9 would clear the bar (score ≥ 7). Real numbers land in your alerts.
One radar across every board
Upwork Freelancer PeoplePerHour LinkedIn Reddit n8n community Hacker News RemoteOK
The pipeline

Three steps, running around the clock.

No feeds to check, no keywords to babysit. AgentScout does the watching so you only see the jobs worth your time.

01 · scan

We watch the wire

Every hour AgentScout sweeps Freelancer, the n8n community jobs board, Reddit, and LinkedIn contract roles — including the niche boards the generic job bots never watch.

02 · score

Claude rates it 1 to 10

Each posting is read in full and scored on budget, stack fit, and recency. Only jobs that clear 7 are worth interrupting you for.

03 · ping

It lands where you work

The alert hits your dashboard — and pushes to Telegram or Slack if you've connected them — with the score, the budget, and a one-tap Apply link, usually within an hour of the job going live.

The score

Every job gets a number, so you can trust the ping.

Claude reads the whole posting and rates it on the things that decide whether a job is actually worth chasing.

  • budget A real project budget, above your floor — not a $50 one-off.
  • stack fit Agents, LLM apps, LangChain, n8n, Claude, MCP, RAG.
  • recency Fresh postings, caught before the flood of applicants.
  • signal A real brief from a real buyer, not a fishing post.
  • personal fit Weighed against your profile and hard rules, once you set them — a great gig outside your stack scores lower for you.

Set a profile (a few lines on your skills and rate) and Check Fit scores each job for you, not just the posting. No profile yet? You still get the gig-quality score above.

9/10
LangGraph multi-agent research pipeline
Freelancer · $4,000 fixed · posted 40m ago
budget $4,000 LangGraph + RAG fixed budget fresh · 40m
Coverage

What it watches

Eight sources, one watchlist of the signals that mean agent work. Tuned so you catch the real thing and skip the noise.

sources
UpworkFreelancerPeoplePerHour LinkedIn contractRedditn8n community Hacker NewsRemoteOK
signals
AI agentsmulti-agentagent orchestration LangChainLangGraphCrewAI LlamaIndexAutoGenMCP RAGvector DBsn8n MakeZapiervoice AI (Vapi/Retell) ClaudeOpenAILLM apps API integrationworkflow automationSemantic Kernel Pydantic AIOpenAI Agents SDKTemporal InngestPineconeWeaviate Qdranttool callingAI SDR conversational AILangSmithprompt engineering + 60 more signals
Delivered your way

The alert reaches you where you already work.

A qualifying job lands the moment it clears the bar — in your web dashboard, as a Telegram message, or mirrored to Slack — with its score, budget, and a one-tap Apply link. Want it? Draft researches the client and writes the proposal, right there.

  • alert Only gigs that clear a 7/10, scored by Claude, so you know it's worth the tap.
  • draft A tailored proposal, client researched, in your voice. Not quite right? Ask for it shorter, more technical, led with a different angle.
  • research A client dossier before you bid: who they are, money signals, red flags, and the angle to lead with.
  • pause Mute the alerts anytime, resume the second you're back on the hunt.

Everything is a one-click button in your web dashboard — and the same tools answer as /commands in Telegram, like the conversation here.

example conversation
AgentScout bot
9/10Freelancer · $4,000 fixed
LangGraph multi-agent research pipeline
fixed budget · posted 40m ago
Apply →
/draft 4581
AgentScout bot

Proposal ready. Client is a seed-stage legal-AI startup scaling agent reliability under lawyer-grade scrutiny.

"I build multi-agent pipelines that hold up in production, evaluation baked in from day one. For your citation-checking flow, I'd start by…"

Copy · Open on Freelancer →
The toolkit

Everything between "job posted" and "contract signed."

Alerts are the start. AgentScout rides along the whole hunt — scoring the jobs you find, researching the buyer, writing the pitch, and keeping score of what wins.

Read-only by design. AgentScout never bids, posts, or clicks as you — auto-bidding bots break marketplace rules and risk your account. It does the homework; you apply as yourself.

Score any job you're looking at

The browser extension puts a Check Fit button on Upwork, Freelancer, PeoplePerHour, and LinkedIn — one click rates the posting against your profile and hard rules (set them once in Settings) before you spend a connect. No extension handy? Paste the posting into Check Fit on your dashboard, or after /score in Telegram.

Automated lead-triage agent (n8n + Claude)
$2,000 fixed · payment verified · 12 hires

We need an automation developer to build an AI agent that triages inbound leads into HubSpot…

Check Fit
9/10
Squarely n8n + Claude agent work. Solid fixed budget, credible client.
n8nClaudeHubSpotwebhooks
Full verdict + one-click Draft → your dashboard

Know the buyer before you bid

Hit Research on any job — one click in your dashboard or /research in Telegram — and AgentScout investigates the client with live web search: who they are, how they pay, the red flags to watch, and the one angle your proposal should lead with. No more bidding blind.

/research 4581
Seed-stage legal-AI startup — small team, recently funded
Payment-verified · five-figure marketplace spend · 12 hires
n8n, Claude API, HubSpot webhooks
Scope is light on success criteria — pin deliverables early
Lead with a citation-checker architecture sketch — prove you get their stack
→ then Draft opens with that exact angle

Your hunt, measured

A private dashboard tracks the funnel that matters: alerts → opened → checked → drafted → won. Report a win in one tap and the scoring learns your taste.

agentscoutjobs.com/me
Sent32
Open14
Fit9
Draft6
Won2
9/10LangGraph multi-agent research pipeline won
8/10Retell voice agent with tool callingdrafted
7/10Migrate Zapier flows to LangChainopened

It learns what you won't take

Set hard rules once in Settings — budget floors, stacks you're done with, deal-breakers. Every Check Fit applies them strictly: a violated rule caps the score and names itself.

your rules
/criteria no WordPress, fixed-price $1k+ only
Rules saved. Every Check Fit now applies them strictly.
2/10 — violates your rule: WordPress work.
the whole kit
scoredraftresearchstatsrulesskillsbudget floorslackbookmarkletextensionpause
Market intelligence

We publish what this market pays. Free.

Every week AgentScout turns its scan data into the State of AI-Agent Jobs: median budgets by platform, which tools are gaining, where the work actually is. The numbers you set your rate with — free for everyone, subscriber or not.

Read the latest report
state of ai-agent jobs · weekly
Median fixed budget$2.4k
Tool momentumn8n
Hottest stackClaude + n8n
Scanned this week1,200+
sample week shown — the live numbers are in the report
How it compares

Not another keyword alert. Not a risky bot.

Generic alert tools fire on keywords and stop there. Auto-bidders apply for you and gamble your account. AgentScout scores, researches, and drafts — and leaves the applying to you.

AgentScout Generic alert bots Auto-bidding tools
Sources watched 8 boards, incl. niche 1–2, keyword feeds Upwork only
Scoring Claude, tuned to AI-agent work Keyword match Basic filters
AI proposal drafts Templated
Client research
Runs on your account Never — read-only Read-only Yes — ban risk
Price Free, or from $8/mo flat $3–27/mo $20–49/mo + credits
Pricing

Simple pricing. Founding price locked for life.

Founding offer: be one of the first 20 subscribers and pay $8/mo — locked forever. After seat 20, Pro costs $15/mo.

Free
$0
The daily pulse, on the house.
  • Daily digest, top 3 scored picks
  • Weekly AI-agent market report
Join free
Pro
$15/mo · or $144/yr
Every gig, scored. Swept hourly. Save $36/yr.
  • Real-time alerts, every scan — dashboard, Telegram, Slack
  • Check Fit — score any job (extension, bookmarklet, dashboard)
  • One-click proposal drafts + client research dossiers
  • Personal dashboard, win-rate stats
Opens after founding sells out
First 20 seats
Founding
$8/mo · locked for life
Everything in Pro. You pay $8/mo instead of $15 — and that price never rises.
  • Everything in Pro
  • Every starter kit included$79 each otherwise, current and future
  • Your price never rises, ever
  • Founding badge + a direct line to the roadmap
Lock founding price

Two ways in: the invite-reviewed free beta (full Pro free while the beta runs, no card; beta testers get first claim on the founding rate) — or subscribe now and lock $8/mo for life. Or refer 2 builders and your founding seat is on the house.

Questions

Good to know

How do alerts reach me?

Every alert lands in your personal web dashboard, and pushes to Telegram and Slack in real time if you connect them. Join right here — your welcome email includes your dashboard link and a one-tap Telegram connect.

How fast are alerts?

The scanner sweeps every hour. We optimize for signal, not for being first: you get the handful of gigs worth your time, not every posting the moment it lands.

What counts as a match?

Anything squarely AI-agent or automation, agents, LLM apps, Claude/OpenAI, n8n, Make, Zapier, LangChain, RAG, MCP, scoring 7 or higher.

Freelance, contract, or remote?

All three — freelance and contract gigs plus remote roles, including remote full-time. Onsite-only, location-locked jobs are filtered out, so everything you get can be done from anywhere.

Can I pause?

Yes — one click in your dashboard's settings, or /pause in Telegram. Resume the same way.

How do proposal drafts work?

Draft researches the client and writes a tailored proposal for that exact job, yours to send in one tap — one click on any alert in your dashboard, or /draft in Telegram. Add notes to redraft: shorter, more technical, lead with X.

Will it auto-apply for me?

No — deliberately. Marketplace rules ban automated bidding, and bots that do it put your account at risk. AgentScout finds, scores, researches, and writes; you apply, as yourself.

What's Check Fit?

Score any job you're already looking at, against your profile and hard rules. One click via the browser extension or bookmarklet, paste it into your dashboard, or send /score in Telegram.

Get started

Free to start. The next $5k gig is worth catching first.

Join the free beta