Perplexity Launches Computer for Counsel: A Multi-Model Agentic Layer for Legal Workflows

Illustrative simulation for engineers. Not related to stay authorized knowledge or actual shopper issues.</p>
</div>

<div class=”section-label”>1 · Choose a authorized workflow</div>
<div class=”tabs” id=”tabs”>
<div class=”tab energetic” data-k=”nda”>
<div class=”t”>Third-party NDA consumption</div>
<div class=”d”>Review purple flags, fill entities, route for signature.</div>
</div>
<div class=”tab” data-k=”reg”>
<div class=”t”>Regulatory monitoring</div>
<div class=”d”>Build a dashboard of US state privateness & adtech legal guidelines.</div>
</div>
<div class=”tab” data-k=”case”>
<div class=”t”>Case analysis + citations</div>
<div class=”d”>Research non-compete enforceability, export cited PDF.</div>
</div>
</div>

<div class=”controls”>
<button class=”act run” id=”run”> Run agentic workflow</button>
<button class=”act reset” id=”reset”>Reset</button>
</div>

<div class=”progress”><div class=”bar” id=”bar”></div></div>
<div class=”section-label”>2 · Pipeline execution</div>
<div class=”pipe” id=”pipe”></div>

<div class=”out” id=”out”>
<h3 id=”outTitle”>Deliverable</h3>
<div class=”ship” id=”outBody”></div>
</div>

<div class=”foot”>
<span class=”src”>Facts sourced from Perplexity’s “Introducing Computer for Counsel” (Jun 24, 2026).</span>
<span class=”model”>Marktechpost</span>
</div>
</div>

<script>
// —- Illustrative workflow definitions (info drawn from Perplexity’s launch weblog) —-
var FLOWS = {
nda: {
title: “Clean NDA bundle, routed for signature”,
steps: [
{n:”Plan task”, model:null, meta:”planner”, body:”Decompose into: extract parties → flag risky clauses → normalize template → route for e-signature.”},
{n:”Route models”, model:[“Reasoning model”,”Extraction model”], meta:”router · 20+ fashions”, physique:”Router assigns a reasoning mannequin to clause evaluation and an extraction mannequin to entity seize.”},
{n:”Call MCP connectors”, mannequin:null, conn:[“Docusign”,”NetDocuments”], meta:”MCP”, physique:”Pulls the counterparty draft from WebDocuments; prepares the Docusign envelope and signatory fields.”},
{n:”Synthesize”, mannequin:[“Reasoning model”], meta:”synthesis”, physique:”Flags an uncapped indemnity and a lacking residuals clause. Fills entity and signatory data. Prepares a clear copy.”},
{n:”Verify & route”, mannequin:null, meta:”human-in-loop”, physique:”Each flag hyperlinks to the supply clause. Lawyer approves; Computer routes for signature through Docusign.”}
],
ship: ‘Reviewed NDA with 2 purple flags surfaced <a category=”cite” title=”Illustrative supply hyperlink”>[clause 4.2]</a> <a category=”cite” title=”Illustrative supply hyperlink”>[clause 7.1]</a>, entities auto-filled, and a clear copy queued in Docusign for approval. The lawyer makes the ultimate name earlier than something is shipped.’
},
reg: {
title: “Shareable US state privateness & adtech dashboard”,
steps: [
{n:”Plan task”, model:null, meta:”planner”, body:”Decompose into: enumerate states → check law status → gather cases → build shareable dashboard.”},
{n:”Route models”, model:[“Research model”,”Reasoning model”], meta:”router · 20+ fashions”, physique:”Router makes use of a analysis mannequin for jurisdiction sweep and a reasoning mannequin for standing classification.”},
{n:”Call MCP connectors”, mannequin:null, conn:[“Midpage”,”Open web”], meta:”MCP”, physique:”Queries Midpage for statutes and related instances; cross-checks efficient dates on the open internet.”},
{n:”Synthesize”, mannequin:[“Reasoning model”], meta:”synthesis”, physique:”Marks which states have legal guidelines in impact vs pending. Cites Midpage for every holding.”},
{n:”Publish dashboard”, mannequin:null, meta:”export”, physique:”Generates a shareable dashboard. Every standing cell hyperlinks to its statute or case.”}
],
ship: ‘A shareable dashboard exhibiting which US states have privateness/adtech legal guidelines in impact vs pending, every cell citing Midpage <a category=”cite” title=”Illustrative supply hyperlink”>[statute]</a> <a category=”cite” title=”Illustrative supply hyperlink”>[case]</a>. Built to be re-run as legal guidelines change.’
},
case: {
title: “Non-compete enforceability memo (PDF, cited)”,
steps: [
{n:”Plan task”, model:null, meta:”planner”, body:”Decompose into: scope question → retrieve cases → run citator → summarize → export PDF.”},
{n:”Route models”, model:[“Research model”,”Reasoning model”], meta:”router · 20+ fashions”, physique:”Research mannequin gathers post-2024 FTC-ban instances; reasoning mannequin weighs enforceability.”},
{n:”Call MCP connectors”, mannequin:null, conn:[“Midpage citator”], meta:”MCP”, physique:”Midpage citator checks whether or not every cited case remains to be good legislation.”},
{n:”Synthesize”, mannequin:[“Reasoning model”], meta:”synthesis”, physique:”Summarizes key instances. Flags unsettled splits throughout circuits. Notes the place legislation is in flux.”},
{n:”Export PDF”, mannequin:null, meta:”export”, physique:”Exports a PDF memo. Every declare hyperlinks to its underlying authority.”}
],
ship: ‘A PDF memo on non-compete enforceability after the FTCu2019s 2024 ban: key instances summarized, unsettled splits flagged <a category=”cite” title=”Illustrative supply hyperlink”>[case A]</a> <a category=”cite” title=”Illustrative supply hyperlink”>[case B]</a>, every verified by the Midpage citator earlier than export.’
}
};

var energetic = “nda”, operating = false, timers = [];

operate postHeight(){
attempt{
var h = doc.getElementById(‘root’).offsetHeight + 40;
mum or dad.postMessage({sort:’cfc-resize’, peak:h}, ‘*’);
}catch(e){}
}

operate clearTimers(){ timers.forEach(clearTimeout); timers = []; }

operate buildSteps(){
var pipe = doc.getElementById(‘pipe’);
pipe.innerHTML = ”;
FLOWS[active].steps.forEach(operate(s,i){
var el = doc.createElement(‘div’);
el.className = ‘step’; el.id = ‘step’+i;
var chips = ”;
if(s.mannequin){ s.mannequin.forEach(operate(m){ chips += ‘<span class=”chip mannequin”>’+m+'</span>’; }); }
if(s.conn){ s.conn.forEach(operate(c){ chips += ‘<span class=”chip”>’+c+'</span>’; }); }
el.innerHTML =
‘<div class=”row”><span class=”dot”></span><span class=”title”>’+(i+1)+’. ‘+s.n+'</span>’+
‘<span class=”meta”>’+s.meta+'</span></div>’+
‘<div class=”physique”>’+s.physique+'</div>’+
(chips?(‘<div class=”chips”>’+chips+'</div>’):”);
pipe.appendChild(el);
});
}

operate setTab(okay){
if(operating) return;
energetic = okay;
doc.querySelectorAll(‘.tab’).forEach(operate(t){
t.classList.toggle(‘energetic’, t.dataset.okay===okay);
});
reset();
}

operate reset(){
clearTimers(); operating = false;
doc.getElementById(‘run’).disabled = false;
doc.getElementById(‘bar’).model.width = ‘0’;
doc.getElementById(‘out’).classList.take away(‘present’);
buildSteps();
setTimeout(postHeight, 50);
}

operate run(){
if(operating) return;
operating = true;
doc.getElementById(‘run’).disabled = true;
doc.getElementById(‘out’).classList.take away(‘present’);
var steps = FLOWS[active].steps;
var n = steps.size;
steps.forEach(operate(s,i){
// reveal
timers.push(setTimeout(operate(){
var el = doc.getElementById(‘step’+i);
el.classList.add(‘present’,’operating’);
doc.getElementById(‘bar’).model.width = ((i+0.5)/n*100)+’%’;
postHeight();
}, i*900 + 150));
// full
timers.push(setTimeout(operate(){
var el = doc.getElementById(‘step’+i);
el.classList.take away(‘operating’); el.classList.add(‘finished’);
doc.getElementById(‘bar’).model.width = ((i+1)/n*100)+’%’;
postHeight();
}, i*900 + 750));
});
// remaining output
timers.push(setTimeout(operate(){
doc.getElementById(‘outTitle’).textContent = ‘✓ ‘ + FLOWS[active].title;
doc.getElementById(‘outBody’).innerHTML = FLOWS[active].ship +
‘<br><br><span model=”colour:#76B900;font-size:11px;”>Judgment and technique stick with the lawyer. Enterprise knowledge shouldn’t be used for coaching.</span>’;
doc.getElementById(‘out’).classList.add(‘present’);
doc.getElementById(‘run’).disabled = false;
operating = false;
postHeight();
}, n*900 + 400));
}

doc.getElementById(‘tabs’).addEventListener(‘click on’, operate(e){
var t = e.goal.closest(‘.tab’); if(t) setTab(t.dataset.okay);
});
doc.getElementById(‘run’).addEventListener(‘click on’, run);
doc.getElementById(‘reset’).addEventListener(‘click on’, reset);

buildSteps();
window.addEventListener(‘load’, operate(){ setTimeout(postHeight, 80); });
window.addEventListener(‘resize’, postHeight);
setTimeout(postHeight, 200);
</script>
</physique>
</html>
”
model=”width:100%;border:0;show:block;peak:600px;overflow:hidden;”
loading=”lazy” title=”Computer for Counsel — Agentic Workflow Simulator”>

# Perplexity's cited search primitive — the inspiration Computer for Counsel builds on.
# This is the general public Sonar API, not the Computer for Counsel product itself.
import os, requests

resp = requests.publish(
    "https://api.perplexity.ai/chat/completions",
    headers={"Authorization": f"Bearer {os.environ['PERPLEXITY_API_KEY']}"},
    json={
        "mannequin": "sonar-pro",
        "messages": [
            {"role": "user",
             "content": "Is a non-compete signed in California enforceable in 2026?"}
        ],
        # Restrict grounding to trusted sources, the way in which a lawyer would.
        "search_domain_filter": ["law.cornell.edu", "courtlistener.com", "ca.gov"],
    },
    timeout=60,
)
knowledge = resp.json()
print(knowledge["choices"][0]["message"]["content"])   # the reply
for url in knowledge.get("citations", []):              # each supply, for verification
    print("supply:", url)