Modern newsrooms operate under unprecedented pressure: breaking news cycles measured in minutes, shrinking editorial budgets, and audiences demanding personalized content across multiple platforms. Chinese LLMs like DeepSeek V4, GLM-4, and Qwen3 can help journalists work faster and smarter by automating routine tasks, analyzing large document sets, and adapting content for diverse audiences. TokenEase's unified API provides news organizations with cost-effective access to these models.
Transform press releases, earnings reports, and data feeds into structured news briefs with lead paragraphs and key facts.
import requests
def generate_news_brief(source_material, publication_style, target_length):
response = requests.post(
"https://tokenease.io/v1/chat/completions",
headers={"Authorization": "Bearer YOUR_TOKENEASE_API_KEY"},
json={
"model": "deepseek-v4",
"messages": [
{"role": "system", "content": f"You are a journalist writing for {publication_style}. Generate news briefs with strong leads, accurate facts, and neutral tone. Follow AP style guidelines."},
{"role": "user", "content": f"Length: {target_length}\nSource material:\n{source_material}\n\nWrite news brief."}
]
}
)
return response.json()["choices"][0]["message"]["content"]
source = """
Apple Inc. (AAPL) today reported quarterly revenue of $94.9 billion, up 6% year over year.
Earnings per share: $1.64 vs analyst estimate of $1.60. iPhone revenue: $45.2B. Services revenue: $24.2B, up 14%.
CEO Tim Cook: "We are excited about Apple Intelligence and its potential to transform the user experience."
Guidance: Q4 revenue expected between $93B-$96B."""
brief = generate_news_brief(source, "business wire service", "200 words")
Analyze large sets of leaked documents, FOIA responses, or court filings to identify patterns, key figures, and story angles.
def analyze_document_dump(documents_summary, investigation_focus, known_entities):
response = requests.post(
"https://tokenease.io/v1/chat/completions",
headers={"Authorization": "Bearer YOUR_TOKENEASE_API_KEY"},
json={
"model": "qwen3-235b",
"messages": [
{"role": "system", "content": "Analyze document collections for investigative journalism. Identify patterns, connect entities, flag inconsistencies, and suggest story angles. Maintain source skepticism."},
{"role": "user", "content": f"Focus: {investigation_focus}\nKnown entities: {known_entities}\nDocuments:\n{documents_summary}\n\nAnalyze and suggest angles."}
]
}
)
return response.json()["choices"][0]["message"]["content"]
docs = """
Doc 1: Internal memo from CFO to board, dated 2024-03-15, mentions "revenue recognition adjustments"
Doc 2: Email chain between accounting manager and external auditor, questions about Q4 timing
Doc 3: Board minutes 2024-06-20: "Discussed restatement possibility, deferred decision"
Doc 4: Whistleblower complaint: "Revenue booked before delivery on 47 transactions"
Doc 5: External audit report draft: "Material weakness in revenue controls"
"""
focus = "Potential accounting fraud or earnings management"
entities = "CFO John Smith, Accounting Manager Lisa Chen, External Auditor Deloitte"
angles = analyze_document_dump(docs, focus, entities)
Extract key quotes, identify contradictions, and summarize themes from interview transcripts.
def analyze_interview(transcript, interview_subject, story_angle):
response = requests.post(
"https://tokenease.io/v1/chat/completions",
headers={"Authorization": "Bearer YOUR_TOKENEASE_API_KEY"},
json={
"model": "glm-4-plus",
"messages": [
{"role": "system", "content": "Analyze interview transcripts for journalism. Extract key quotes, identify newsworthy statements, flag contradictions with public records, and summarize themes."},
{"role": "user", "content": f"Subject: {interview_subject}\nAngle: {story_angle}\nTranscript:\n{transcript}\n\nAnalyze and extract."}
]
}
)
return response.json()["choices"][0]["message"]["content"]
transcript = """
Q: When did you first learn about the contamination?
A: I think it was sometime in early 2024. Maybe March or April.
Q: Did you report it to regulators?
A: We followed all internal protocols. I'm not sure about external reporting.
Q: Your 2023 annual report states "full environmental compliance." Was that accurate?
A: We believed it was accurate at the time based on the information available.
Q: Three former employees say they warned management in 2022.
A: I don't recall receiving those warnings.
"""
analysis = analyze_interview(transcript, "CEO of manufacturing company", "environmental cover-up investigation")
Generate multiple headline variants optimized for different platforms and A/B test performance predictions.
def optimize_headlines(article_summary, target_platforms, keywords):
response = requests.post(
"https://tokenease.io/v1/chat/completions",
headers={"Authorization": "Bearer YOUR_TOKENEASE_API_KEY"},
json={
"model": "kimi-k2",
"messages": [
{"role": "system", "content": "Generate platform-optimized headlines. Balance clickability with accuracy. Include SEO keywords naturally. Avoid clickbait while maximizing engagement."},
{"role": "user", "content": f"Keywords: {keywords}\nPlatforms: {target_platforms}\nArticle:\n{article_summary}\n\nGenerate headlines for each platform."}
]
}
)
return response.json()["choices"][0]["message"]["content"]
summary = "Federal Reserve signals potential interest rate cuts in September due to cooling inflation and slowing job growth. Markets rallied on the news."
platforms = ["Website homepage", "Twitter/X", "Google News", "Newsletter subject line"]
keywords = "Federal Reserve, interest rates, inflation, September"
headlines = optimize_headlines(summary, platforms, keywords)
Adapt news stories for international audiences with culturally aware translation and local context addition.
def adapt_for_international(original_article, target_market, local_context_needed):
response = requests.post(
"https://tokenease.io/v1/chat/completions",
headers={"Authorization": "Bearer YOUR_TOKENEASE_API_KEY"},
json={
"model": "deepseek-v4",
"messages": [
{"role": "system", "content": f"Adapt news articles for {target_market} audiences. Add local context, explain cultural references, and adjust framing for regional relevance."},
{"role": "user", "content": f"Local context needed: {local_context_needed}\nMarket: {target_market}\nArticle:\n{original_article}\n\nAdapt for local audience."}
]
}
)
return response.json()["choices"][0]["message"]["content"]
article = """
US tech companies face new AI regulation under the proposed federal framework.
The bill would require safety testing for models above certain compute thresholds
and mandate disclosure of training data sources."
"""
context = "Explain how this compares to EU AI Act and China's algorithmic recommendation regulations"
adapted = adapt_for_international(article, "Asian business readers", context)
Analyze reader comments to identify trending topics, sentiment shifts, and potential follow-up stories.
def analyze_audience_engagement(comments_sample, article_topic, engagement_metrics):
response = requests.post(
"https://tokenease.io/v1/chat/completions",
headers={"Authorization": "Bearer YOUR_TOKENEASE_API_KEY"},
json={
"model": "qwen3-235b",
"messages": [
{"role": "system", "content": "Analyze audience engagement data. Identify trending subtopics, sentiment patterns, misinformation to correct, and potential follow-up story angles."},
{"role": "user", "content": f"Metrics: {engagement_metrics}\nTopic: {article_topic}\nComments:\n{comments_sample}\n\nAnalyze engagement and suggest follow-ups."}
]
}
)
return response.json()["choices"][0]["message"]["content"]
sample = """
1. "What about the environmental impact? The article didn't mention that."
2. "This is just corporate PR. Where's the critical analysis?"
3. "My city tried this and it failed. Here are the details..."
4. "Can you do a comparison with European approaches?"
5. "The cost figures seem off. I work in this industry and..."
"""
metrics = "12K comments, 45K shares, 89% positive sentiment, top question: environmental impact"
topic = "New infrastructure project announcement"
engagement = analyze_audience_engagement(sample, topic, metrics)
Get $1 free credits (1M tokens) to automate routine reporting and analyze document sets.
Start with TokenEase →