Meta tags:
description=Manage fleets of local and cloud agents from one surface. Plan, delegate, review, and ship without leaving your editor.;
keywords=AI coding agent, coding AI, AI software engineer, SWE agent, AI coding assistant, cloud agent, parallel agent;

Headings (most frequently used words):

the, devin, desktop, agents, of, you, it, all, models, tab, team, forevery, engineer, power, an, ide, exactly, when, need, on, acp, teams, building, with, make, your, own, so, good, can, work, without, learn, about, our, plans, trusted, by, developers, frequently, asked, questions, one, space, for, every, agent, ship, instant, codebase, context, never, miss, detail, free, world, class, effortless, handoff, to, cloud,

Text of the page (most frequently used words):
and (64), the (62), for (30), config (26), devin (22), your (19), #desktop (19), self (19), agents (17), agent (17), server (14), with (14), from (13), you (11), every (10), download (9), plan (9), mcp (9), context (9), windsurf (8), more (8), all (8), our (8), manage (8), working (8), ago (8), learning (8), rate (8), ide (7), work (7), local (7), window_sizes (7), value (7), update (6), one (6), spaces (6), language (6), that (6), across (6), run (6), engine (6), embeddings (6), learn (5), pricing (5), enterprise (5), center (5), full (5), code (5), pattern (5), n_layer (5), gradient (5), its (5), norm (5), mlp (5), add (5), support (4), will (4), over (4), read (4), new (4), command (4), customers (4), teams (4), search (4), fast (4), cognition (4), engineering (4), cloud (4), implement (4), build (4), model (4), class (4), test (4), char (4), _compute_window_sizes (4), def (4), __init__ (4), waiting (4), onboarding (4), gpt (4), tuning (4), get (3), jetbrains (3), what (3), same (3), settings (3), fully (3), are (3), place (3), multi (3), developers (3), select (3), per (3), access (3), create (3), retrieve (3), completion (3), type (3), diagnostics (3), extension (3), data (3), comments (3), list (3), tools (3), design (3), engineers (3), them (3), shared (3), research (3), engineer (3), scratch (3), done (3), world (3), models (3), review (3), faketimers (3), zig (3), ship (3), return (3), char_to_window (3), range (3), long_window (3), train (3), autoresearch (3), flow (3), micrograd (3), graph (3), each (3), reverse (3), output (3), n_embd (3), establish (3), training (3), baseline (3), home (3), privacy (2), docs (2), stay (2), exactly (2), plans (2), change (2), extensions (2), workflows (2), intact (2), when (2), only (2), name (2), can (2), how (2), building (2), view (2), while (2), experience (2), accessible (2), trusted (2), contact (2), month (2), free (2), request (2), sign (2), visit (2), macos (2), without (2), issues (2), documentation (2), logs (2), traces (2), projects (2), vercel (2), navigation (2), checking (2), python (2), issue (2), status (2), files (2), codebase (2), workspace (2), rust (2), linear (2), fix (2), typescript (2), messages (2), team (2), already (2), partner (2), multiple (2), day (2), first (2), lead (2), nvidia (2), excited (2), share (2), harvey (2), built (2), background (2), running (2), product (2), now (2), laptop (2), instead (2), starting (2), single (2), ramp (2), use (2), into (2), sales (2), swe (2), coding (2), before (2), detail (2), 185 (2), 187 (2), informational (2), line (2), setfaketimermarker (2), order (2), bug (2), false (2), acp (2), supercomplete (2), next (2), tab (2), 131 (2), append (2), 130 (2), len (2), 129 (2), 128 (2), 127 (2), short_window (2), 126 (2), 125 (2), 124 (2), 123 (2), 122 (2), 121 (2), 120 (2), 119 (2), super (2), 118 (2), 117 (2), space (2), utf (2), autograd (2), backward (2), reused (2), expressions (2), pytorch (2), test_engine (2), node (2), parents (2), forward (2), then (2), backprop (2), client (2), protocol (2), head_dim (2), torch (2), parameter (2), vocab_size (2), block (2), embedding (2), wte (2), transformer (2), window_size (2), cos_sin (2), 42m (2), switch (2), activation (2), gelu (2), tune (2), wsd (2), warmdown (2), ratio (2), ready (2), warmup (2), phase (2), increase (2), muon (2), matrix (2), widen (2), hidden (2), dimension (2), sessions (2), editor (2), twitter, linkedin, choices, terms, service, policy, intellij, idea, pycharm, webstorm, continues, available, happening, current, including, legacy, does, progress, remain, migrated, branding, changing, lose, anything, arrives, standard, air, carry, automatically, also, latest, version, page, upgrade, announcement, foundation, introduce, managing, kanban, management, front, know, remains, frequently, asked, questions, startups, agencies, enterprises, 4000, million, worldwide, users, stats, let, talk, seat, 200, max, popular, pro, about, overview, demo, good, show, jira, confluence, atlassian, telemetry, insights, incidents, monitors, dashboards, datadog, deployments, analyze, official, refactoring, gopls, products, subscriptions, invoices, payouts, refunds, stripe, intellisense, pyright, stack, filter, sentry, clangd, nodes, images, components, styles, webhooks, figma, opinionated, formatter, enforces, consistent, style, prettier, pages, databases, notion, definition, inline, analyzer, query, initiatives, cycles, find, problems, javascript, eslint, channels, send, user, profiles, slack, extend, skills, plugins, uses, make, own, senior, architect, ciprian, nechita, gives, intelligent, but, permissions, flexibility, their, machines, development, benefits, faster, hands, environment, natural, fit, snappier, fits, way, lot, today, member, technical, staff, rahul, chalamala, been, closely, tool, lets, together, subhash, ranjan, joining, preview, complex, help, define, they, coordinate, joey, wang, internal, spectre, long, efforts, carrying, organizational, legal, seamlessly, collaborate, custom, extends, humans, understanding, shaiyon, hariri, makes, easy, dispatch, monitor, array, bring, making, easier, jump, between, tasks, preserve, financial, modal, designed, close, effortless, handoff, dashboard, real, time, store, stream, events, kafka, websockets, render, onto, three, globe, unlimited, fastest, rapidly, deeply, diff, push, never, miss, comment, inconsistent, implementation, 179, 189, silently, ignores, errors, 25869, investigate, dependency, assumes, clean, state, mark, flags, 188, sets, clock, resetting, finds, exact, lines, needs, milliseconds, instant, where, reconciliation, scheduled, predicts, thought, not, just, edit, sequence_len, upper, window_pattern, git, worktrees, encoding, approval, improve, unicode, normalization, 12m, authentication, error, api, endpoints, feat, redesign, cascade, opencode, claude, codex, written, topo, sorts, applies, accumulating, sub, sum, correctly, check, matches, pytest, sanity, gradients, against, small, expression, variable, accumulation, actually, exercised, pass, itself, topologically, seed, stored, closures, accumulated, reached, returns, stashes, _backward, closure, chain, rule, pushing, inputs, records, computation, backpropagates, mode, autodiff, empty, repo, scaffold, edges, minimal, scalar, like, powered, n_kv_head, kv_dim, 138, n_head, 137, 136, zeros, x0_lambdas, 135, ones, resid_lambdas, 134, bias, lm_head, 133, 132, modulelist, moduledict, module, provides, flop, estimation, utilities, _precompute_rotary_embeddings, init_weights, handles, rotary, windowed, attention, patterns, builds, layer, head, implements, defined, attn, 116, includes, syntax, highlighting, autocomplete, debugging, trace, debug, need, power, decide, write, chase, edge, cases, best, platform, jsx, 231, col, screen, reader, optimized, launchpad, main, clear, filters, pull, display, board, session, fleets, surface, delegate, leaving, log, resources, solutions,

Text of the page (random words):
devin desktop devin product solutions customers resources pricing contact sales log in download the home for every agent you run learn more devin desktop manage fleets of local and cloud agents from one surface plan delegate review and ship without leaving your editor download for macos visit on desktop to download sign up the home for every agent you run learn more agent editor new session sessions spaces widen the model hidden dimension working switch the mlp activation to gelu 42m ago learning rate tuning increase the muon matrix learning rate working add a learning rate warmup phase pr is ready tune the wsd warmdown ratio waiting for ci establish the training baseline 3h ago add qk norm and value embeddings 1h ago sessions board list display status space pull request agent clear filters running 2 widen the model hidden dimension working learning rate tuning increase the muon matrix learning rate working waiting for review 2 learning rate tuning add a learning rate warmup phase pr is ready learning rate tuning tune the wsd warmdown ratio waiting for ci done 3 switch the mlp activation to gelu 42m ago establish the training baseline 3h ago add qk norm and value embeddings 1h ago main launchpad 0 0 screen reader optimized ln 231 col 29 spaces 2 utf 8 typescript jsx cognition platform enterprise windsurf settings agent command center a team of agents for every engineer devin desktop is the home for coding agents to do your best work you decide what to build then your agents write the code chase the edge cases and test every detail autoresearch establish the training baseline add qk norm and value embeddings a world class ide the power of an ide exactly when you need it read trace and debug every change your agents ship devin desktop includes a full ide with syntax highlighting autocomplete and debugging tools built in for you to stay in flow autoresearch train py 116 self mlp mlp config 117 118 def forward self x ve cos_sin window_size 119 x x self attn norm x ve cos_sin window_size 120 x x self mlp norm x 121 return x 122 123 124 class gpt class gpt type gpt go to gpt implements the full transformer language model defined in train py builds wte embedding block layer list and output head handles value embeddings rotary embeddings and windowed attention patterns provides init_weights _precompute_rotary_embeddings and flop estimation utilities nn module 125 def __init__ self config 126 super __init__ 127 self config config 128 self window_sizes self _compute_window_sizes config 129 self transformer nn moduledict 130 wte nn embedding config vocab_size config n_embd 131 h nn modulelist block config i for i in range config n_layer 132 133 self lm_head nn linear config n_embd config vocab_size bias false 134 self resid_lambdas nn parameter torch ones config n_layer 135 self x0_lambdas nn parameter torch zeros config n_layer 136 value embeddings 137 head_dim config n_embd config n_head 138 kv_dim config n_kv_head head_dim agent client protocol all the models all the agents work across models and agents powered by the agent client protocol acp micrograd implement backprop implement backprop from scratch a minimal scalar autograd engine like micrograd i ll build a single value type that records the computation graph as expressions run then backpropagates with reverse mode autodiff starting from an empty repo with the node scaffold the data its gradient and the edges to its parents engine py 9 each forward op returns a new value and stashes a local _backward closure the chain rule for that op pushing the output gradient into its inputs engine py 17 engine py 13 now the backward pass itself topologically order the graph from the output seed its gradient to 1 and run the stored closures in reverse so each node is fully accumulated before its parents are reached engine py 12 sanity checking the gradients against pytorch on a small expression with a reused variable so the accumulation is actually exercised test_engine py 9 python m pytest test_engine py q done a 50 line autograd engine written from scratch backward topo sorts the graph and applies each op s local gradient in reverse accumulating with so reused sub expressions sum correctly the gradient check matches pytorch to 1e 6 code swe 1 6 fast devin local agents devin cloud devin local codex claude agent opencode cascade local micrograd spaces spaces onboarding flow redesign feat onboarding flow design build onboarding ui working update onboarding api endpoints working fix authentication error 12m ago improve unicode normalization waiting for approval implement utf 8 encoding 2d ago one space for every agent use spaces to share context and git worktrees across all your agents supercomplete autoresearch train py 117 def __init__ self config 118 super __init__ 119 self config config 120 self window_sizes self _compute_window_sizes config self window_sizes self _compute_window_sizes config 121 122 def _compute_window_sizes self config 123 pattern config window_pattern upper 124 long_window config sequence_len 125 short_window long_window 2 126 char_to_window l long_window 0 s short_window 0 127 128 for i in range config n_layer for i in range config n_layer 129 char pattern i len pattern char pattern i len pattern 130 window_sizes append char_to_window char window_sizes append char_to_window char 131 return window_sizes return window_sizes tab tab ship supercomplete predicts your next thought not just your next edit agents on acp fast context where is reconciliation scheduled instant codebase context fast context finds the exact files and lines your agent needs in milliseconds 1 bug setfaketimermarker sets clock to false instead of resetting it bug faketimers zig 188 3 flags mark all as read test order dependency first test assumes clean state investigate 25869 test ts 20 23 setfaketimermarker silently ignores errors informational faketimers zig 179 189 comment at line 185 187 is inconsistent with implementation informational faketimers zig 185 187 never miss a detail rapidly or deeply review every agent diff before you push free world class models unlimited access to swe 1 6 the fastest coding model in the world plan md to build a dashboard for real time store sales data we will stream events from kafka over websockets and render them onto a three js globe view plan implement in cloud effortless handoff to the cloud the only ide designed for you to close your laptop customers teams building with devin desktop ramp harvey nvidia modal intact financial devin desktop makes it easy to dispatch and monitor our array of agents from a single command center we re excited to partner with cognition to bring the agents ramp engineers already use into one shared workspace making it easier to jump between tasks preserve context and get more done shaiyon hariri research engineer at harvey we built our internal background agent spectre to work across long running engineering efforts while carrying organizational context for our legal research engineering product and design teams to seamlessly collaborate with devin desktop s support for custom background agents that context now extends to every engineer s laptop so humans and agents work from the same shared understanding instead of starting from scratch joey wang engineering lead nvidia is joining cognition s research preview for multi agent support in devin desktop our engineers run multiple agents across complex workflows every day and we re excited to help define how they share context and coordinate in one place subhash ranjan engineering lead ai tools we ve been working closely with cognition as a design partner on multi agent support in devin desktop our engineers run multiple agents every day and devin desktop is the first tool that lets them manage all of them together with shared context from one place rahul chalamala member of technical staff devin desktop gives our teams the same intelligent agent experience but with the full permissions and flexibility of their local machines for development work that benefits from a faster more hands on environment it s a natural fit it s snappier it s accessible and it fits the way a lot of our developers are working today ciprian nechita senior it architect make devin desktop your own extend devin with the tools skills and plugins your team already uses slack mcp server search channels and messages send and read messages and access user profiles eslint extension find and fix problems in your javascript and typescript code linear mcp server list create update and query issues projects initiatives cycles and comments rust analyzer language server code completion go to definition and inline diagnostics for rust notion mcp server retrieve and manage pages databases and comments search across your workspace prettier extension opinionated code formatter that enforces a consistent style across your codebase figma mcp server get files nodes and images manage comments components styles and webhooks clangd language server c and c language server with completion navigation and diagnostics sentry mcp server retrieve issue data and stack traces search filter and update issue status windsurf pyright extension fast type checking intellisense and diagnostics for python stripe mcp server create and manage customers products subscriptions invoices payouts and refunds gopls language server the official go language server for completion navigation and refactoring vercel mcp server manage projects and deployments analyze logs and search vercel documentation datadog mcp server retrieve telemetry insights and manage incidents monitors logs dashboards and traces atlassian mcp server access jira and confluence manage issues and create enterprise documentation show more so good you can t work without it download for macos visit on desktop to download sign up request a demo pricing overview learn about our plans free 0 download pro popular 20 per month select plan max new 200 per month select plan teams 80 mo 40 mo per full seat select plan enterprise let s talk contact us stats trusted by developers 1m users trusted by over a million developers worldwide 4000 enterprise customers startups agencies and enterprises frequently asked questions what is devin desktop devin desktop is the new name for windsurf we re building on the ide foundation of windsurf to introduce the command center for managing all your agents in one place the agent command center spaces kanban view and multi agent management is front and center while the full ide experience you know remains fully accessible read the announcement how do i upgrade to devin desktop from windsurf devin desktop arrives as a standard over the air update so your plan pricing extensions and settings all carry over automatically you can also download the latest devin desktop version from the download page will i lose anything if i update the ide your extensions workflows settings and in progress work will all remain intact and will be fully migrated when you update only the name and branding are changing learn more in the docs does my plan or pricing change no your current plan and pricing stay exactly the same including legacy windsurf enterprise plans learn more in the docs what is happening to jetbrains support windsurf for jetbrains intellij idea pycharm webstorm and more continues to be available for download get windsurf for jetbrains privacy policy terms of service your privacy choices linkedin x twitter