API Reference

Send a message and receive an AI-generated response. The conversation history is maintained server-side.

Request Body

Response

{
  "response": "Quantum computing uses qubits..."
}

Return the full conversation history for the current session.

Response

{
  "history": [
    {"role": "user", "content": "Hello"},
    {"role": "assistant", "content": "Hi! How can I help?"}
  ]
}

Clear the conversation history and start fresh.

Response

{"message": "Conversation history cleared"}

Update the system prompt that guides the AI's behaviour.

Request Body

Response

{"message": "System prompt updated", "prompt": "You are a pirate..."}

Send audio, get transcription + AI text response + audio response. Full STT → Chat → TTS pipeline.

Request

multipart/form-data with a file field containing the audio file.

Response

{
  "transcription": "What is AI?",
  "response": "AI stands for...",
  "audio": "<base64-encoded MP3>"
}

Transcribe an audio file to text using OpenAI Whisper.

Request

multipart/form-data with a file field.

Response

{"text": "Hello, how are you?"}

Convert text to MP3 audio using OpenAI TTS.

Request Body

Response

Binary MP3 audio (audio/mpeg).

Send text, get AI text response + audio response (Chat → TTS).

Request Body

Response

{
  "response": "Here is my answer...",
  "audio": "<base64-encoded MP3>"
}

Upload video → extract audio → AI chat → text + audio response.

Request

multipart/form-data with a file field containing the video.

Upload video → extract audio → transcribe to text with Whisper.

Send text → chat → TTS → text + audio response.

Request Body

Upload video → extract a frame at 1 second → describe with AI Vision.

Generate code from a natural-language description.

Request Body

Review code and return feedback on bugs, security issues, and improvements.

Request Body

Explain what a piece of code does in plain language.

Request Body

Refactor code for better quality, readability, and maintainability.

Request Body

Submit a task for the AI agent to execute autonomously using tool calling.

Request Body

Response

{
  "result": "The answer is 15.",
  "tools_available": ["calculator", "datetime_now", "text_transform", "json_extract"]
}

Return the agent's conversation and tool-use history.

Clear the agent's conversation history.

Describe an uploaded image using OpenAI Vision.

Request

multipart/form-data with a file field.

Response

{"description": "A sunset over a mountain range..."}

Answer a specific question about an uploaded image.

Request

multipart/form-data with file and question fields.

Generate an image from a text prompt using DALL-E 3.

Request Body

Response

{"url": "https://oaidalleapi...", "prompt": "A cat wearing a top hat"}

Generate a Word document (.docx) from text. Returns binary file download.

Request Body

Generate a PowerPoint presentation (.pptx) from text. Long text is automatically split across slides.

Generate a PDF document (.pdf) from text.

Natural-language database query with row-level access control per role.

Request Body

Response

{"answer": "There were 1,247 orders placed in February."}

Return the database schema (tables and columns).

Execute a raw SELECT or SHOW query. Admin/company role only.

Request Body

Clear the DB chatbot conversation history.

Select a working database after customer login.

List companies the user belongs to.

Select a company database by company number.

Translate text to a target language.

Request Body

Response

{"translated": "Bonjour le monde", "target_language": "French"}

Summarize text with configurable style and length.

Request Body

Response

{"summary": "A concise summary of the input text..."}

Analyse text sentiment — returns sentiment label, confidence score, and explanation.

Request Body

Response

{
  "sentiment": "positive",
  "confidence": 0.95,
  "explanation": "Strong positive language detected"
}

Extract structured data from unstructured text given a schema description.

Request Body

Response

{
  "data": {"name": "Alice", "email": "alice@example.com", "phone": "555-1234"}
}

Check text for harmful content using the OpenAI Moderation API.

Request Body

Response

{
  "flagged": false,
  "categories": {},
  "category_scores": {"violence": 0.0012, "hate": 0.0003}
}

Generate an embedding vector for a text string.

Request Body

Response

{
  "embedding": [0.012, -0.034, 0.056, ...],
  "dimensions": 1536
}

Compute the cosine similarity between two texts (0 to 1).

Request Body

Response

{"text1": "cat", "text2": "kitten", "score": 0.92}

Fetch a URL and summarize its content, or answer a specific question about it.

Request Body

Response

{
  "url": "https://example.com/article",
  "summary": "The article discusses..."
}

Store a memory with optional metadata. An embedding vector is generated automatically for semantic recall.

Request Body

Response

{
  "id": 0,
  "content": "User prefers dark mode",
  "metadata": {"category": "preferences"}
}

Recall the most relevant memories by semantic similarity to a query.

Request Body

Response

{
  "memories": [
    {"id": 0, "content": "User prefers dark mode", "metadata": {}, "score": 0.92}
  ],
  "count": 1
}

Clear all stored memories.

Response

{"message": "All memories cleared"}

Ingest plain text into a knowledge base namespace. Text is automatically chunked and embedded.

Request Body

Response

{"namespace": "my-app", "chunks_added": 3, "source": "faq.md"}

Fetch a URL, extract text, chunk, embed, and store into a knowledge base. SSRF-protected.

Request Body

Response

{"namespace": "my-app", "chunks_added": 12, "source": "https://example.com/faq"}

Query a knowledge base with a question. Uses RAG to find relevant chunks and generate an answer.

Request Body

Response

{"answer": "Returns are accepted within 30 days.", "sources": ["faq.md"], "namespace": "my-app"}

List all knowledge base namespaces with their chunk counts.

Response

{"knowledge_bases": [{"namespace": "my-app", "chunks": 15}]}

Clear all documents from a knowledge base namespace.

Request Body

Response

{"message": "Knowledge base 'my-app' cleared."}

Translate text between African languages (Swahili, Yoruba, Hausa, Amharic, Wolof, Lingala, Zulu, and more).

Request Body

Response

{"translated": "Habari, u hali gani?", "target_language": "Swahili"}

Detect which African language a text is written in.

Request Body

Response

{"language": "Swahili", "confidence": 0.95, "script": "Latin", "region": "East Africa"}

List all supported African languages.

Response

{"languages": ["Swahili", "Yoruba", "Hausa", "Amharic", "Wolof", "Lingala", "Zulu", "..."]}

Transcribe audio with African language support. Send audio as multipart form data with an optional language hint (ISO-639-1).

Form Fields

Response

{"text": "Habari yako"}

Convert African language text to speech. Returns audio/mpeg bytes.

Request Body

Chat with an AI knowledgeable about African geography, history, cultures, economics, and current affairs.

Request Body

Response

{"response": "Kenya's main exports include tea, coffee, horticultural products, and textiles..."}

Clear the African context chat conversation history.

Response

{"message": "African chat history cleared."}

Moderate content in African languages, detecting hate speech, tribalism, and misinformation.

Request Body

Response

{"flagged": false, "reason": "", "language_detected": "Swahili", "categories": {"hate_speech": false, "tribalism": false, "misinformation": false, "violence": false, "discrimination": false}}

Format an amount in an African currency (CFA, Naira, Shilling, Rand, Cedi, etc.).

Request Body

Response

{"formatted": "KSh 1,500.50", "currency_code": "KES", "currency_name": "Kenyan Shilling", "amount": 1500.5}

Estimate conversion between African currencies.

Request Body

Response

{"source_amount": 1000, "source_currency": "KES", "target_amount": 3150, "target_currency": "NGN", "rate": 3.15, "note": "Approximate rate"}

List all supported African currencies.

Response

{"currencies": {"KES": "Kenyan Shilling", "NGN": "Nigerian Naira", "ZAR": "South African Rand", "...": "..."}}

Ask a question about mobile money services (M-Pesa, MTN MoMo, Orange Money, Airtel Money, EcoCash).

Request Body

Response

{"response": "To send money via M-Pesa, dial *334# and follow the prompts..."}

Parse a mobile money transaction message or USSD code into structured data.

Request Body

Response

{"provider": "M-Pesa", "transaction_type": "send", "amount": 500, "currency": "KES", "recipient": "+254712345678", "reference": "QH45G7Z2", "status": "completed", "raw_text": "Confirmed. Ksh500 sent to +254712345678"}

Ask about African farming, crops, livestock, and agricultural practices.

Request Body

Response

{"response": "Cassava is one of the most important crops in Sub-Saharan Africa..."}

Ask education and exam prep questions for African educational systems (WAEC, JAMB, KCSE, Baccalauréat).

Request Body

Response

{"response": "In the WAEC Biology syllabus, photosynthesis is..."}

Ask about African business law and regulatory guidance.

Request Body

Response

{"response": "In Nigeria, the Companies and Allied Matters Act (CAMA) requires..."}

Get an African proverb by theme or culture.

Request Body

Response

{"proverb": "If you want to go fast, go alone...", "language": "Igbo", "culture": "Igbo", "meaning": "...", "theme": "unity"}

Tell an African folktale, legend, or story based on a topic or character.

Request Body

Response

{"story": "Long ago, in the land of the Ashanti..."}

Stream a chat response token by token. Returns text/event-stream.

Request Body

Response (SSE stream)

data: Once
data: upon
data: a
data: time
data: [DONE]

Send multiple messages and receive individual responses for each.

Request Body

Response

{
  "results": ["Answer to first...", "Answer to second..."],
  "count": 2
}

Customer login — validates against customers_credential + customers, requires is_verified = 'yes'.

Request Body

Response

{
  "access_token": "eyJhbGciOi...",
  "token_type": "bearer",
  "role": "customer",
  "firstname": "Alice"
}

Admin login — requires company_employee_links.is_admin = 1 and status = 1.

Employee login — requires company_employee_links.status = 1. JWT includes allow_tables_list.

Company login — requires customers.type = 'webcompany'. Same access rights as admin.

Get API usage statistics for the currently authenticated user.

Health check — returns API status and version. No authentication required.

Response

{"status": "ok", "version": "0.1.0"}