Skip to content

Commit 07c14ae

Browse files
BinWang28BinWang28
committed
update docs
1 parent 823b6fb commit 07c14ae

File tree

7 files changed

+265
-17
lines changed

7 files changed

+265
-17
lines changed

‎docs/mkdocs/docs/hle.md‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -2,7 +2,7 @@
22

33
MiroFlow's evaluation on the HLE benchmark demonstrates capabilities in multimodal reasoning and question answering tasks that require human-level understanding across vision and language.
44

5-
More details: [HLE Dataset on HuggingFace](https://huggingface.co/datasets/cais/hle)
5+
More details: [Humanity's Last Exam](https://arxiv.org/abs/2501.14249)
66

77
---
88

‎docs/mkdocs/docs/hle-text-only.md‎ renamed to ‎docs/mkdocs/docs/hle_text_only.md‎

File renamed without changes.

‎docs/mkdocs/docs/index.md‎

Lines changed: 5 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -18,10 +18,11 @@
1818

1919
**Oct 2025** -
2020

21-
- Add support for Index
22-
- Add support for BrowseComp-EN evaluation
23-
- Add support for MiroAPI https://github.com/MiroMindAI/MiroFlow/pull/76
24-
21+
22+
- 📝 Added support for HLE-Text evaluation benchmark [#81](https://github.com/MiroMindAI/MiroFlow/pull/81)
23+
- 🧠 Added support for HLE (Humanity's Last Exam) benchmark [#79](https://github.com/MiroMindAI/MiroFlow/pull/79)
24+
- 🌐 Added support for BrowseComp-EN evaluation benchmark [#78](https://github.com/MiroMindAI/MiroFlow/pull/78)
25+
- 🔌 Added support for MiroAPI integration [#76](https://github.com/MiroMindAI/MiroFlow/pull/76)
2526

2627
- 📊 Added support for FinSearchComp evaluation benchmark [#51](https://github.com/MiroMindAI/MiroFlow/pull/51)
2728
- 🔍 Added support for XBench-DS (Deep Search) evaluation [#47](https://github.com/MiroMindAI/MiroFlow/pull/47)

‎docs/mkdocs/docs/tool_audio.md‎

Lines changed: 90 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,90 @@
1+
# Audio Tools (`tool-audio`)
2+
3+
Audio processing capabilities including transcription and audio-based question answering.
4+
5+
---
6+
7+
## Configuration
8+
9+
```yaml title="Agent Configuration"
10+
main_agent:
11+
tool_config:
12+
- tool-audio
13+
```
14+
15+
**Environment Variables:**
16+
17+
- `OPENAI_API_KEY`: **Required**. OpenAI API key
18+
- `OPENAI_BASE_URL`: API base URL. Default: `https://api.openai.com/v1`
19+
- `OPENAI_TRANSCRIPTION_MODEL_NAME`: Default: `gpt-4o-transcribe`
20+
- `OPENAI_AUDIO_MODEL_NAME`: Default: `gpt-4o-audio-preview`
21+
22+
---
23+
24+
## Function Reference
25+
26+
### `audio_transcription(audio_path_or_url: str)`
27+
28+
Transcribe audio file to text using OpenAI's Whisper models.
29+
30+
**Parameters:**
31+
32+
- `audio_path_or_url`: Local file path or URL
33+
- Supported formats: MP3, WAV, M4A, FLAC, OGG, WebM
34+
- Not supported: E2B sandbox paths, YouTube URLs
35+
36+
**Returns:**
37+
38+
- `str`: Full transcription of the audio file
39+
40+
**Example:**
41+
42+
```python
43+
# Transcribe local audio
44+
transcription = await audio_transcription("/data/meeting.mp3")
45+
46+
# Transcribe from URL
47+
transcription = await audio_transcription("https://example.com/podcast.wav")
48+
```
49+
50+
---
51+
52+
### `audio_question_answering(audio_path_or_url: str, question: str)`
53+
54+
Answer questions based on audio content using GPT-4o Audio.
55+
56+
**Parameters:**
57+
58+
- `audio_path_or_url`: Local file path or URL (same formats as transcription)
59+
- `question`: Question to answer about the audio content
60+
61+
**Returns:**
62+
63+
- `str`: Answer with audio duration information
64+
65+
**Example:**
66+
67+
```python
68+
# Ask about content
69+
answer = await audio_question_answering(
70+
"/data/lecture.mp3",
71+
"What are the main topics discussed?"
72+
)
73+
74+
# Get summary
75+
answer = await audio_question_answering(
76+
"https://example.com/interview.wav",
77+
"Summarize the key points."
78+
)
79+
```
80+
81+
**Important Notes:**
82+
83+
- Cannot access E2B sandbox files (`/home/user/`)
84+
- YouTube URLs not supported (use VQA tools instead)
85+
- Includes audio duration in response
86+
87+
---
88+
89+
!!! info "Documentation Info"
90+
**Last Updated:** October 2025 · **Doc Contributor:** Team @ MiroMind AI

‎docs/mkdocs/docs/tool_reading.md‎

Lines changed: 71 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,71 @@
1+
# Reading Tools (`tool-reading`)
2+
3+
Read and convert various document formats (DOC, PDF, Excel, etc.) to markdown for easy processing.
4+
5+
---
6+
7+
## Configuration
8+
9+
```yaml title="Agent Configuration"
10+
main_agent:
11+
tool_config:
12+
- tool-reading
13+
```
14+
15+
**Environment Variables:**
16+
17+
- `SERPER_API_KEY`: Required for certain operations
18+
- `JINA_API_KEY`: Required for document processing
19+
20+
---
21+
22+
## Function Reference
23+
24+
### `read_file(uri: str)`
25+
26+
Read various types of resources and convert them to markdown format.
27+
28+
**Parameters:**
29+
30+
- `uri`: The URI or path of the resource to read. Supported:
31+
- Local file paths (e.g., `/path/to/document.pdf`)
32+
- `file:` URIs (e.g., `file:/path/to/document.pdf`)
33+
- `http:` / `https:` URLs (will be downloaded automatically)
34+
- `data:` URIs (base64-encoded)
35+
36+
**Supported Formats:**
37+
38+
- Documents: DOC, DOCX, RTF, ODT
39+
- Presentations: PPT, PPTX, ODP
40+
- Spreadsheets: XLS, XLSX, CSV, ODS
41+
- PDFs: PDF documents
42+
- Archives: ZIP files
43+
- Images and text files
44+
45+
**Returns:**
46+
47+
- `str`: Content in markdown format, or error message if reading fails
48+
49+
**Example:**
50+
51+
```python
52+
# Read a local PDF
53+
result = await read_file("file:/path/to/document.pdf")
54+
55+
# Read from URL
56+
result = await read_file("https://example.com/report.pdf")
57+
58+
# Read local file (auto-converted to file: URI)
59+
result = await read_file("/data/spreadsheet.xlsx")
60+
```
61+
62+
**Important Notes:**
63+
64+
- Cannot access E2B sandbox files (`/home/user/`)
65+
- Use local file paths provided in the original instruction
66+
- Downloaded files are automatically cleaned up
67+
68+
---
69+
70+
!!! info "Documentation Info"
71+
**Last Updated:** October 2025 · **Doc Contributor:** Team @ MiroMind AI

‎docs/mkdocs/docs/tool_searching_serper.md‎

Lines changed: 92 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,92 @@
1+
# Searching Tools - Serper (`tool-searching-serper`)
2+
3+
Lightweight Google search and web scraping via Serper API using NPM package.
4+
5+
!!! tip "Which Tool to Use?"
6+
- **`tool-searching-serper`**: Fast Google search + basic scraping (NPM-based)
7+
- **`tool-searching`**: Full-featured with Wikipedia, Archive.org, JINA (Python-based)
8+
9+
---
10+
11+
## Configuration
12+
13+
```yaml title="Agent Configuration"
14+
main_agent:
15+
tool_config:
16+
- tool-searching-serper
17+
```
18+
19+
**Environment Variables:**
20+
21+
- `SERPER_API_KEY`: **Required**. Get at [serper.dev](https://serper.dev)
22+
23+
---
24+
25+
## Function Reference
26+
27+
### `google_search(q: str, gl: str = "us", hl: str = "en", location: str = None, num: int = 10, tbs: str = None, page: int = 1)`
28+
29+
Perform Google searches via Serper API.
30+
31+
**Parameters:**
32+
33+
- `q`: Search query (required)
34+
- `gl`: Country code (e.g., 'us', 'uk', 'cn'). Default: 'us'
35+
- `hl`: Language (e.g., 'en', 'zh', 'es'). Default: 'en'
36+
- `location`: City location (e.g., 'San Francisco, California, United States')
37+
- `num`: Number of results. Default: 10
38+
- `tbs`: Time filter ('qdr:h'=hour, 'qdr:d'=day, 'qdr:w'=week, 'qdr:m'=month, 'qdr:y'=year)
39+
- `page`: Page number. Default: 1
40+
41+
**Returns:**
42+
43+
- `str`: JSON formatted search results
44+
45+
**Example:**
46+
47+
```python
48+
# Basic search
49+
results = await google_search("artificial intelligence")
50+
51+
# With filters
52+
results = await google_search("latest news", tbs="qdr:d", num=20)
53+
```
54+
55+
---
56+
57+
### `scrape(url: str)`
58+
59+
Scrape website content using Serper.
60+
61+
**Parameters:**
62+
63+
- `url`: Website URL to scrape
64+
65+
**Returns:**
66+
67+
- `str`: Scraped content
68+
69+
**Example:**
70+
71+
```python
72+
content = await scrape("https://example.com/article")
73+
```
74+
75+
---
76+
77+
## Comparison: Serper vs Full Searching
78+
79+
| Feature | `tool-searching-serper` | `tool-searching` |
80+
|---------|------------------------|------------------|
81+
| Google Search | ✅ | ✅ |
82+
| Web Scraping | ✅ Basic | ✅ Advanced |
83+
| Wikipedia | ❌ | ✅ |
84+
| Archive.org | ❌ | ✅ |
85+
| YouTube Info | ❌ | ✅ |
86+
| Speed | ⚡ Faster | Slightly slower |
87+
| Dependencies | Node.js/NPM | Python only |
88+
89+
---
90+
91+
!!! info "Documentation Info"
92+
**Last Updated:** October 2025 · **Doc Contributor:** Team @ MiroMind AI

‎docs/mkdocs/mkdocs.yml‎

Lines changed: 6 additions & 12 deletions
Original file line numberDiff line numberDiff line change
@@ -65,16 +65,7 @@ nav:
6565
- xBench-DeepSearch: xbench_ds.md
6666
- FinSearchComp: finsearchcomp.md
6767
- HLE: hle.md
68-
- HLE(text only): hle_text_only.md
69-
70-
# - Benchmarks:
71-
# - GAIA-Validation-Text-Only: gaia_validation_text_only.md
72-
# - GAIA-Test: gaia_test.md
73-
# - BrowseComp-EN: browsecomp_en.md
74-
# - FutureX: futurex.md
75-
# - xBench-DeepSearch: xbench_ds.md
76-
# - FinSearchComp: finsearchcomp.md
77-
# - Download Datasets: download_datasets.md
68+
- HLE-Text-Only: hle_text_only.md
7869

7970

8071

@@ -85,12 +76,15 @@ nav:
8576
- "": ""
8677
- "": ""
8778
- "": ""
79+
- tool-reading: tool_reading.md
80+
- tool-searching: tool_searching.md
81+
- tool-searching-serper: tool_searching_serper.md
82+
- tool-audio: tool_audio.md
83+
- tool-audio-os: tool_audio_os.md
8884
- tool-reasoning: tool_reasoning.md
8985
- tool-reasoning-os: tool_reasoning_os.md
9086
- tool-image-video: tool_vqa.md
9187
- tool-image-video-os: tool_vqa_os.md
92-
- tool-audio-os: tool_audio_os.md
93-
- tool-searching: tool_searching.md
9488
- tool-python: tool_python.md
9589
- "": ""
9690
- "": ""

0 commit comments

Comments
 (0)