Every chat app, driven from one command.

Ask ChatGPT, Claude, Gemini, DeepSeek, Grok and more in your own logged-in Chrome, from the terminal or from your agent. One provider or all of them at once. Your agent gets the reply, not the page.

terminalinstall
you@mac ~ %PS C:\Users\you>you@ubuntu:~$ npm install -g ai-browser-bridge
v0.7.15macOS, Windows, LinuxNode 24No API keysMIT
Chrome · ~/.ai-browser-bridge/chrome-profilebridge ask · 3 providers
claude.aidriven by bridge
Which Postgres index fits a 40M-row time series table?
Claude

It depends mostly on two things: how rows arrive, and what your queries look like. With 40M rows you have two realistic choices, BRIN and B-tree, and many setups end up using both.

BRIN fits when data is inserted roughly in time order (append-only logs, sensor readings, events). BRIN stores only the min and max timestamp for each block range. The index ends up kilobytes to a few MB, compared with roughly 1GB or more for a B-tree on 40M rows, and it costs almost nothing on inserts. It’s great for wide range scans like “all rows from last week” or “aggregate March”. It’s poor for very selective lookups, because it can only skip blocks, not find rows.

 sql CREATE INDEX ON readings USING brin (ts) WITH (pages_per_range = 32);

Message ClaudeWhich Postgres index fits a 40M-row time series table?
claude
chat.deepseek.comdriven by bridge
Which Postgres index fits a 40M-row time series table?
DeepSeek

For a 40M-row time series table, the best index depends entirely on how the data is physically ordered on disk. BRIN is the ideal choice if rows are appended in time order (e.g., an INSERT-only log table), while B-tree is the safer general-purpose choice if physical order is not guaranteed or your queries need precise, low-latency point lookups.

🥇 BRIN (Block Range Index): The Time-Series Specialist

If your 40M rows are stored roughly in chronological order (common for append-only time series), BRIN is dramatically more efficient than a standard index.

Message DeepSeekWhich Postgres index fits a 40M-row time series table?
deepseek
grok.comdriven by bridge
Which Postgres index fits a 40M-row time series table?
GrokThought for 12s

BRIN on the timestamp column.

For a 40M-row time series table with mostly chronological (append-only) inserts, a BRIN index is usually the best fit:

SQL CREATE INDEX CONCURRENTLY idx_ts_brin ON your_table USING brin (ts) WITH (pages_per_range = 64, autosummarize = on); Why BRIN fits Extremely small (often KB–low MB vs hundreds of MB or more for a B-tree) Low write and maintenance cost Strong for time-range queries (ts BETWEEN ... or ts > ...)

Message GrokWhich Postgres index fits a 40M-row time series table?
grok
terminalyour agent · ~/shop-api
you@mac ~ %PS C:\Users\you>you@ubuntu:~$ bridge ask --provider claude,deepseek,grok --json "Which Postgres index fits a 40M-row time series table?" sending to 3 chat apps in your Chrome ✓ claude ok · reply returned · 47.7s ✓ deepseek ok · reply returned · 39.3s ✓ grok ok · reply returned · 56.6s 3 of 3 replied in 147.0s · the pages stayed in Chrome
context metertokens
Page HTML485,839
Reply text1,195
kept out of your agent484,644
10chat apps behind one command
34MCP tools for agents
3permission modes for repo access
0API keys. Your own accounts and plans.

Log in once. Ask from anywhere.

The bridge opens Chrome on its own profile. You sign in to each chat app one time, and every repo and every agent reuses it.

Start the bridge Chrome

Opens a Chrome window on a separate profile. Sign in to the chat apps you use. The sessions stay there for every repo.

bridge chrome start
~/.ai-browser-bridge/chrome-profile
ChatGPTsigned in
Claudesigned in
Geminisigned in
DeepSeeksigned in
Groksigned in
Perplexitysigned in

Ask from the terminal or your agent

Name one provider or a list to fan out. Agents connect over MCP and call the same ask tool.

bridge serve
terminalterminal
# one provider
you@mac ~ %PS C:\Users\you>you@ubuntu:~$ bridge ask --provider claude --json "Review this migration"

# fan out to three
you@mac ~ %PS C:\Users\you>you@ubuntu:~$ bridge ask --provider claude,deepseek,grok --json "same question, three models"

# give Claude Code the bridge as MCP tools
you@mac ~ %PS C:\Users\you>you@ubuntu:~$ claude mcp add ai-browser-bridge -- bridge serve

Get clean JSON back

Each provider returns its reply text as JSON, ready for a script or an agent to read.

--json
terminalstdout
[
  { "provider": "claude", "ok": true, "text": "It depends mostly on…" },
  { "provider": "deepseek", "ok": true, "text": "For a 40M-row time…" },
  { "provider": "grok", "ok": true, "text": "BRIN on the timestamp…" }
]

What the bridge does, running

Short loops of real commands. Each one plays while it is on screen.

Sort conversations into Projects

Give it a plan file. It moves each conversation, checks that it landed, and slows down when ChatGPT says too many requests.

Search your conversations

Search titles and message content from the terminal. Matches come back as JSON.

Ask several providers at once

List providers after --provider and get every reply back, keyed by provider.

Permission modes and checkpoints

Pick read-only, ask or auto. Every patch gets a checkpoint you can restore.

Mention files with @

Type @ to pick a repo file. It is expanded into the prompt before the chat sees it.

Generate video with Flow

Send a prompt to Google Flow, then download the clips into .bridge/ in your repo.

Your agent reads the answer, not the page.

A chat page is thousands of lines of markup, scripts and old messages. The bridge reads the reply inside Chrome and hands back only the text, so your agent keeps its context for your code.

what a page scrape would send161,814tokens of page HTML
what bridge returns477tokens of reply text
It depends mostly on two things: how rows arrive, and what your queries look like. With 40M rows you have two realistic choices, BRIN and B-tree, and many setups end up using both. BRIN fits when data is inserted roughly in time order (append-only logs, sensor readings, events). BRIN stores only the min and max timestamp for each block range. The index ends up kilobytes to a few MB, compared with roughly 1GB or more for a B-tree on 40M rows, and it costs almost nothing on inserts. It’s great for wide range scans like “all rows from last week” or “aggregate March”. It’s poor for very selective lookups, because it can only skip blocks, not find rows.  sql CREATE INDEX ON readings USING brin (ts) WITH (pages_per_range = 32); Before you rely on it, check that the physical order really matches time order:  sql SELECT correlation FROM pg_stats WHERE tablename = 'readings' AND attname = 'ts'; If the result is close to 1.0, BRIN will work well. If it’s much lower (backfills, late-arriving data, heavy updates), try the brin_timestamp_minmax_multi_ops operator class (PG14+), which handles some out-of-order data. Otherwise use a B-tree. B-tree fits when queries are selective or ordered: ORDER BY ts DESC LIMIT 100 (“latest N”) Narrow windows, like the last 5 minutes Filtering by an entity plus time. This is the most common real pattern, and it needs a composite index with the equality column first:  sql CREATE INDEX ON readings (device_id, ts DESC); -- optionally make it covering: -- CREATE INDEX ON readings (device_id, ts DESC) INCLUDE (value); At 40M rows and growing, also consider partitioning. Declarative range partitioning by month or day keeps each partition’s B-tree small and lets the planner skip whole partitions. Retention becomes a quick DROP/DETACH PARTITION instead of a huge DELETE. If you want this handled for you (plus compression), TimescaleDB is built for exactly this workload. Rule of thumb: Append-only data with mostly range or aggregate queries:
Page HTML161,814
Reply text477
claudereply read inside Chromestep 1
bridgereturned the text as JSONstep 2
agentgot the answer, not the pagestep 3

MCP in both directions.

Your agent can call the chat apps. ChatGPT, Claude and Grok can read your repo through a small set of checked tools, never a shell.

bridge serve

signature
ask({ prompt?, providers?, tasks?, timeoutSeconds?, maxConcurrency?, limit?, offset?, maxReplyChars? })

Drive web chats: one prompt fanned across providers, or a `tasks` array of independent Conversations run in parallel (new or resumed). Returns an ordered, paginated result — one row per task with its reply and reopenable Conversation id/url.

register once
claude mcp add ai-browser-bridge -- bridge serve

42 tools in all. Design tools that delete need confirm: true.

bridge ask --tools

Search the repository using ripgrep. Locate symbols, imports, routes, tests, configs, and references.

read-onlygrep, read and git_diff
askwrite and test tools are blocked
autoadds apply_patch and run_tests

Calls reach your Mac through a Cloudflare quick tunnel. apply_patch snapshots each file first, so you can restore it.

Questions

Do I need API keys?

No. The bridge drives the chat apps in your logged-in Chrome, so every request runs on your own account and plan.

What does it run on?

macOS, Windows and Linux, with Node 24.11 or newer and Google Chrome.

What if a chat app changes its page?

Each provider keeps its selectors in one browser layer, so a fix stays in one place.

Can a chat app run commands on my computer?

No. There is no shell. Repo tools are a fixed set, checked and kept inside the repo, under the permission mode you pick.

Where is my data kept?

On your computer. The Chrome profile lives in ~/.ai-browser-bridge/chrome-profile and repo state in .bridge/ inside the repo.

Put every chat app one command away.

Install the CLI, sign in once, and ask from your terminal or your agent.

terminalinstall
you@mac ~ %PS C:\Users\you>you@ubuntu:~$ npm install -g ai-browser-bridge