feat: add web tools config to configure

This commit is contained in:
Peter Steinberger
2026-01-15 05:08:51 +00:00
parent f6a72ef3c2
commit 042b65dfcc
15 changed files with 122 additions and 15 deletions

View File

@@ -576,7 +576,9 @@ The Gateway watches the config and supports hotreload:
### How do I enable web search (and web fetch)?
`web_fetch` works without an API key. `web_search` requires a Brave Search API
key (`BRAVE_API_KEY` or `tools.web.search.apiKey`).
key. **Recommended:** run `clawdbot configure --section web` to store it in
`tools.web.search.apiKey`. Environment alternative: set `BRAVE_API_KEY` for the
Gateway process.
```json5
{
@@ -584,7 +586,7 @@ key (`BRAVE_API_KEY` or `tools.web.search.apiKey`).
web: {
search: {
enabled: true,
apiKey: "BRAVE_API_KEY_HERE", // or set BRAVE_API_KEY
apiKey: "BRAVE_API_KEY_HERE",
maxResults: 5
},
fetch: {
@@ -598,6 +600,7 @@ key (`BRAVE_API_KEY` or `tools.web.search.apiKey`).
Notes:
- If you use allowlists, add `web_search`/`web_fetch` or `group:web`.
- In sandboxed sessions, `web_fetch` auto-enables unless explicitly disabled.
- Daemons read env vars from `~/.clawdbot/.env` (or the service environment).
Docs: [Web tools](/tools/web).