pi-web-providers 3.2.0 → 3.4.0

package/README.md

@@ -66,7 +66,7 @@ reported when the tool is actually called.
 | **Brave**      |   ✔    |          |   ✔    |    ✔     | `BRAVE_SEARCH_API_KEY` / `BRAVE_ANSWERS_API_KEY` |
 | **Cloudflare** |        |    ✔     |        |          | `CLOUDFLARE_API_TOKEN` + `CLOUDFLARE_ACCOUNT_ID` |
 | **Exa**        |   ✔    |    ✔     |   ✔    |    ✔     | `EXA_API_KEY`                                    |
-| **Firecrawl**  |   ✔    |    ✔     |        |          | `FIRECRAWL_API_KEY`                              |
+| **Firecrawl**  |   ✔    |    ✔     |   ✔    |          | `FIRECRAWL_API_KEY`                              |
 | **Gemini**     |   ✔    |          |   ✔    |    ✔     | `GOOGLE_API_KEY`                                 |
 | **Linkup**     |   ✔    |    ✔     |        |    ✔     | `LINKUP_API_KEY`                                 |
 | **Ollama**     |   ✔    |    ✔     |        |          | `OLLAMA_API_KEY`                                 |
@@ -188,7 +188,22 @@ in tool calls.
 Unlike the other managed tools, `web_research` does not accept local timeout,
 retry, polling, or resume controls. Research has one opinionated execution
 style: pi starts it asynchronously, tracks it locally, and saves the final
-report under `.pi/artifacts/research/`.
+report under `.pi/artifacts/research/`. Saved research reports are Markdown
+files with YAML frontmatter for job metadata such as query, provider, status,
+and timestamps.
+
+Use `/web-research` in interactive pi sessions to browse and manage researches.
+The command opens a table of running and finished researches with status, date,
+provider, duration, and title. Running researches show a live spinner and
+elapsed time; press `c` twice to cancel one. Cancellation aborts the provider
+call and records a durable `cancelled` artifact under
+`.pi/artifacts/research/`. Press `Enter` on a finished research to read the
+report in a scrollable Markdown overlay; from there, `c` copies the report as
+Markdown to the clipboard and `i` injects it into the current conversation so
+the agent can build on earlier research.
+
+While researches run, a one-line summary above the editor shows the number of
+active researches with their status and elapsed time.
 
 </details>
 
@@ -315,13 +330,19 @@ scope, or account ID is usually wrong.
 <summary><strong>Firecrawl</strong></summary>
 
 - SDK: `@mendable/firecrawl-js`
-- Supports `web_search` and `web_contents`
+- Supports `web_search`, `web_contents`, and page-scoped `web_answer`
 - Search can optionally include Firecrawl scrape-backed result enrichment
 - Contents extraction uses Firecrawl scrape with markdown-first defaults
+- Answers use Firecrawl scrape's `question` format against one explicit page URL;
+  set `options.url` in the `web_answer` call or
+  `providers.firecrawl.options.answer.url` as a default
 - Exposes search options such as `lang`, `country`, `sources`, `categories`,
   `location`, `timeout`, and `scrapeOptions`
 - Exposes contents options such as `formats`, `onlyMainContent`, `includeTags`,
   `excludeTags`, `waitFor`, `headers`, `location`, `mobile`, and `proxy`
+- Exposes answer options `url`, `onlyMainContent`, `includeTags`,
+  `excludeTags`, `waitFor`, `headers`, `location`, `mobile`, and `proxy`
+- Firecrawl charges 5 credits per page for the `question` format
 - Optional `baseUrl` overrides are supported for self-hosted Firecrawl
   instances, proxies, and testing. API keys are required for Firecrawl Cloud,
   but can be omitted for self-hosted endpoints that do not enforce