Show HN: Jev-browse – browser sub-tasks for coding agents at ~1/3 the cost

Developer danielnc released jev-browse, an MIT-licensed port of browser-use/jev-ultrafast that adds fast_run helpers to browser-harness so coding agents like Claude Code and Codex can hand off an entire browser sub-task in one call. Each step is decided by TypeSafe Jev, a small model answering typed questions in a few hundred milliseconds, and in the author's runs form and navigation tasks were 1.6–3.0× faster and 2.1–5.3× cheaper than the agent driving browser-harness alone at the same pass rate. The tool requires Chrome, browser-harness, Python 3.11+, and a TypeSafe API key, and hands control back to the agent on unknown values, risky clicks, frames, uploads, or sensitive fields.

Fast, typed browser sub-tasks for coding agents. jev-browse adds a few helpers to browser-harness https://github.com/browser-use/browser-harness so that Claude Code or Codex can hand off a whole website sub-task in one call: "search Lisbon, set these two filters, open Casa Flora". Each step is decided by TypeSafe https://docs.typesafe.ai Jev, a small model that answers typed questions in a few hundred milliseconds. The agent then gets the outcome, or a typed reason why it stopped. inside a browser-harness script r = fast run "https://en.wikipedia.org/wiki/Main Page", "Open the Wikipedia article on Gödel's incompleteness theorems", run id="demo-1" print r.status, r.reason, r.url claimed done None https://en.wikipedia.org/wiki/G%C3%B6del%27s ... jev close r.target id Why. When a coding agent drives a browser itself, each click costs a model turn that re-reads its whole context. A five-field form becomes fifteen turns. fast run does the clicking and typing in one call, with one Jev request per decision, and hands back to the agent only when it must: a value it doesn't know, a risky click, a frame or upload it can't handle, or anything sensitive. In the author's runs that made form and navigation tasks 1.6–3.0× faster and 2.1–5.3× cheaper than the agent driving browser-harness alone, with the same pass rate details benchmark . It is a port of browser-use/jev-ultrafast https://github.com/browser-use/jev-ultrafast MIT , turned into helpers any agent can call from a browser-harness script. You need Chrome, browser-harness https://github.com/browser-use/browser-harness connected to it, Python 3.11+, and a TypeSafe API key from console.typesafe.ai/keys https://console.typesafe.ai/keys . Paste this into Claude Code or Codex: Install jev-browse from https://github.com/danielnc/jev-browse by following its install.md: clone it, run its installer, store my TypeSafe API key in the browser-harness agent-workspace .env ask me for it; never print it , and run python3 -m jev browse doctor until it passes. Ask me which text backend I want default: my Claude subscription if the claude CLI is installed and whether to add the jev-browse pointer to my global agent instructions. git clone https://github.com/danielnc/jev-browse ~/jev-browse && cd ~/jev-browse python3 -m jev browse install adds the helpers to browser-harness and links the skill ENV=~/.config/browser-harness/agent-workspace/.env printf 'TYPESAFE API KEY=%s\n' '<your key ' "$ENV" && chmod 600 "$ENV" python3 -m jev browse doctor checks everything and prints what is active browser-harness <<'PY' r = fast run "https://en.wikipedia.org/wiki/Main Page", "Open the Wikipedia article about the Eiffel Tower", run id="hello-1" print r.status, r.url jev close r.target id PY Then add the global pointer tell-your-agent-about-it to your agent's instructions. Without it, agents rarely think to use jev-browse on their own. your agent Claude Code / Codex │ writes one browser-harness script: r = fast run url, goal, values={...} ▼ browser-harness ──CDP──► Chrome: a new background tab owned by jev-browse │ │ │ ┌───────────────────────┘ │ ▼ │ loop: snapshot the page snapshot.js: visible text, controls, fields; sensitive values never read │ → one TypeSafe Jev request: next operation? which target? which value? ──► api.typesafe.ai │ → value missing from the goal? ask the text backend once optional ──► Claude / Codex / │ → safety gates: commit verbs, sensitive fields, frames, uploads, hosts Ollama / any │ → click / type / select / scroll via CDP OpenAI-compatible │ until DONE, BLOCKED, or a typed hand-back confirm required, in frame, text value unavailable, … ▼ RunResult status, reason, url, evidence, trace + a JEV BROWSE RESULT={...} line to verify in the same script - One decision = one small Jev request typically a few hundred milliseconds . No screenshots and no large model in the loop. - Values come from your values= , from the goal itself Jev picks among candidates taken from the goal , or, only after a miss, from a text backend. A grounding gate checks the backend's answer before anything is typed. - It never guesses past its limits. Too many options, frames, shadow DOM, canvas, uploads, missing values, sensitive fields, or a click that would send, pay, delete, or book all hand back with a Reason . The tab stays open for the agent to finish or resume. More: docs/architecture.md https://github.com/danielnc/jev-browse/blob/main/docs/architecture.md . | Use fast run for | Use browser-harness directly for | |---|---| | Multi-step forms and searches fill, pick dates, filter, open a result | Anything visual: charts, layouts, images, "does this look right" | | Link navigation to a known destination | Iframes, shadow DOM, canvas, file uploads, drag and drop | | "Find and click the X" steps jev find / jev click | Sensitive pages banking, health, credentials | | Sub-tasks where you only need the outcome | Pages whose text you should not send to TypeSafe | When fast run hands back, the agent continues with the harness on the same tab, or resumes after supplying what was missing: fast run None, goal, target id=r.target id, values={...} . Zero config beyond TYPESAFE API KEY . Everything else is an environment variable the harness .env counts or an entry in ~/.config/jev-browse/config.toml . The environment wins. The settings you are most likely to change: | Setting config.toml | Environment | Default | |---|---|---| | text.backend | JEV BROWSE TEXT BACKEND | auto : claude if its CLI is installed, else none | | text.fallback | JEV BROWSE TEXT FALLBACK | auto : who answers when a local backend fails | | ollama.url , ollama.model | JEV BROWSE OLLAMA URL , MODEL | unset, qwen3:30b-a3b | | openai.base url , openai.model , openai.api key env | JEV BROWSE OPENAI | unset | | jev.model | JEV BROWSE JEV MODEL | jev-latest | | run.max actions , run.timeout s | JEV BROWSE MAX ACTIONS , JEV BROWSE TIMEOUT S | 30, 90 s | | safety.allowed hosts | JEV BROWSE ALLOWED HOSTS | all hosts | Privacy note on text.fallback . With a local ollama or OpenAI-compatible backend, the default text.fallback = "auto" means: if that backend fails its known-answer check, is unreachable, or returns invalid output twice, the claude CLI answers instead when it is installed , so the goal, field labels, and page excerpt go to Anthropic. If you chose a local model to keep page text on your machine, set text.fallback = "none" JEV BROWSE TEXT FALLBACK=none : the miss then hands back to your agent. doctor prints which fallback is active. All settings, and common setups privacy mode, local model, OpenRouter/Groq/Cerebras/Gemini : docs/configuration.md https://github.com/danielnc/jev-browse/blob/main/docs/configuration.md . python3 -m jev browse config shows what is active and where each value came from. A text backend is asked only when a field's value is implied but not stated "the capital of France" → Paris . | Backend | Text goes to | Cost | Trade-off | |---|---|---|---| | claude default if installed | Anthropic, via your Claude subscription | subscription only; API keys are stripped | a few seconds per miss CLI start-up, mostly hidden by a pre-started process | | codex | OpenAI, via your ChatGPT subscription | subscription only | a few seconds of CLI start-up per miss; not yet benchmarked | | ollama | your own server and the fallback's provider if it fails; see text.fallback | free per token | as fast as your hardware; guarded by a known-answer canary | | openai | any OpenAI-compatible endpoint and the fallback's provider if it fails | provider prices | unmeasured; bring your own model | | none | nowhere | free | misses hand back to the agent privacy mode | Personal fields name, email, phone, address and sensitive fields are never sent to a text backend. Honest numbers and set-up notes: docs/backends.md https://github.com/danielnc/jev-browse/blob/main/docs/backends.md . - What leaves your machine. On every decision, visible page text, element labels, non-personal field values, URL, and title go to TypeSafe api.typesafe.ai . After a miss, the goal, field labels, and up to 2,000 characters of page text go to your text backend , or to its fallback when a local backend fails see text.fallback above . Use it only on pages you are comfortable sending there. Use text backend="none" or text.backend = "none" for sensitive sites. - Sensitive fields passwords, one-time codes, card numbers, CVV, IBAN, national IDs are never typed, never read out of the page, and never sent. Personal fields are reported only as filled or empty. - Commit gate. Clicks that send, pay, delete, book, or confirm need the goal's explicit authorisation, or your confirm= ... in a later script. Otherwise the run hands back confirm required . This is a heuristic verbs plus structural signals such as a confirmation inside a dialog : it lowers the risk of an unwanted click in your signed-in browser. It does not remove it. - Owned tabs. jev-browse works only in background tabs it created or that your new tab created and you adopted . It never touches, focuses, or closes your other tabs. - Hosts. safety.allowed hosts restricts every page jev-browse opens, adopts, or observes. - No telemetry in jev-browse. browser-harness has its own telemetry, which sends script text and helper-call arguments to PostHog when enabled. The installer and doctor warn if it is on. Opt out with browser-harness telemetry disable . - Local data. Run files, traces, and screenshots can contain page text and typed values. They stay in the harness tmp dir and other gitignored paths, and make clean-traces removes them. A skill that never triggers delivers nothing. With only the skill installed, the calling agent used jev-browse in 0 of 5 unprompted tasks in the author's runs. After a short pointer in the global instructions it used it in 2 of 5 , and the explicit fast run call shape saved a turn per task. Paste this into ~/.claude/CLAUDE.md , ~/.codex/AGENTS.md , or your project's agent file: Browser tasks: jev-browse fast path For a multi-step website sub-task search, fill, filter, open a result where you only need the outcome, call jev-browse from a browser-harness script. You don't need to load its skill first: r = fast run url, goal, values={...known field values...}, run id="<unique " print r.status, r.reason, r.detail, r.target id - claimed done is not proof: check the printed JEV BROWSE RESULT line, or js "...", target id=r.target id , in the same script, then jev close r.target id . - On a hand-back r.reason , load the jev-browse skill for what to do next. Resume on the same tab with fast run None, <same goal , target id=r.target id, values={...} ; never re-run from the URL. - Never act on confirm required in the same script: decide first ask me if my request does not clearly cover it , then resume with confirm= ... . - Use browser-harness directly for visual judgement, frames, uploads, and sensitive pages. Page text goes to TypeSafe; pass text backend="none" on sensitive sites. - Close only tab ids that jev-browse returned to you. The same snippet is in docs/global-pointer.md https://github.com/danielnc/jev-browse/blob/main/docs/global-pointer.md . - No iframes, shadow DOM, canvas or visual understanding, file uploads, or pop-up tabs: these hand back. - Pages with very many options or very large state hand back too many options , state too large . - Decisions are text-only. A page whose meaning is visual will confuse it, and it will usually say so visual only , low confidence . - Every decision is a metered TypeSafe request: a few tenths of a cent per task at the time of writing. - The commit gate and the grounding gate are heuristics. Review what a run did r.trace on anything that matters. - The Codex and OpenAI-compatible backends are unmeasured. - Tested with browser-harness 0.1.13 on macOS with Chrome. Other platforms should work but have not been benchmarked. jev-browse was compared with the same coding agent driving browser-harness directly, on public-site navigation, form, and filter tasks, with every outcome verified by code. In the author's runs it was 1.6–3.0× faster and 2.1–5.3× cheaper , with the same pass rate N = 3 each , mostly because the agent needs far fewer turns. Your numbers will differ. Summary: docs/benchmark.md https://github.com/danielnc/jev-browse/blob/main/docs/benchmark.md . Reproduce it, or evaluate your own text backend, with bench/ docs/benchmarking.md https://github.com/danielnc/jev-browse/blob/main/docs/benchmarking.md . - install.md https://github.com/danielnc/jev-browse/blob/main/install.md : step-by-step install, written for an agent to follow - docs/architecture.md https://github.com/danielnc/jev-browse/blob/main/docs/architecture.md : the decision loop, modules, and safety model - docs/configuration.md https://github.com/danielnc/jev-browse/blob/main/docs/configuration.md : every setting - docs/backends.md https://github.com/danielnc/jev-browse/blob/main/docs/backends.md : text backends and their trade-offs - docs/benchmarking.md https://github.com/danielnc/jev-browse/blob/main/docs/benchmarking.md : running the benchmark and text eval - skill/SKILL.md https://github.com/danielnc/jev-browse/blob/main/skill/SKILL.md and skill/reference.md https://github.com/danielnc/jev-browse/blob/main/skill/reference.md : what the agent reads See CONTRIBUTING.md https://github.com/danielnc/jev-browse/blob/main/CONTRIBUTING.md . Development is offline by default: uv sync && make check runs the linter and more than 400 tests with TypeSafe, the CLIs, and Chrome mocked. MIT, see LICENSE https://github.com/danielnc/jev-browse/blob/main/LICENSE . jev-browse ports code from browser-use/jev-ultrafast https://github.com/browser-use/jev-ultrafast MIT, Copyright c 2026 Browser Use ; see NOTICE https://github.com/danielnc/jev-browse/blob/main/NOTICE . Ported files carry a header comment naming the upstream file.