Troubleshooting Guide

April 19, 2026 ยท View on GitHub

This guide helps you diagnose and resolve common issues with ccflare.

Table of Contents

  1. OAuth Authentication Problems
  2. Rate Limiting Issues
  3. Connection Problems
  4. Performance Issues
  5. Account Management Issues
  6. Configuration Problems
  7. Database Issues
  8. Streaming and Analytics Issues
  9. Logging and Debugging
  10. Common Error Messages
  11. Environment Variables Reference
  12. FAQ
  13. Getting Help

OAuth Authentication Problems

Invalid Tokens

Symptom: Requests fail with 401 Unauthorized errors

Error Message: Failed to refresh access token

Solutions:

  1. Check if the access token has expired:

    ccflare --list

    Look for accounts with expired tokens (expires_at in the past)

  2. Refresh the token manually:

    • Remove and re-add the account: 
      ccflare --remove <account-name>
ccflare --add-account <account-name>

  3. Verify the refresh token is still valid in your Anthropic console

Expired Tokens

Symptom: Token refresh attempts fail repeatedly

Error Message: Token expired or missing for account: [name]

Solutions:

  1. ccflare automatically attempts to refresh expired tokens
  2. If automatic refresh fails, re-authenticate the account
  3. Check for refresh token stampede prevention - multiple simultaneous refresh attempts are prevented

PKCE Failures

Symptom: OAuth authorization fails during account setup

Error Messages:

  • Token exchange failed: [error]
  • Invalid code_verifier
  • Refresh promise not found for account

Solutions:

  1. Ensure you're using the complete authorization code including the state fragment (format: code#state)
  2. Don't modify or truncate the authorization code
  3. Complete the OAuth flow within the time limit (codes expire quickly)
  4. Try the authorization flow again from the beginning
  5. Ensure only one refresh attempt happens at a time (refresh stampede prevention is active)

Token Refresh Failures

Symptom: Automatic token refresh fails

Error Messages:

  • Failed to refresh access token
  • Token refresh failed: [error]

Solutions:

  1. Check if the refresh token was revoked in your Anthropic console
  2. Verify the CLIENT_ID environment variable matches your OAuth app
  3. Remove and re-add the account: 
    ccflare --remove <account-name>
ccflare --add-account <account-name>
  4. Check for multiple simultaneous refresh attempts in logs

Rate Limiting Issues

Identifying Rate Limits

ccflare detects rate limits through response headers and HTTP status codes:

  1. Rate Limited Responses:
    • HTTP 429 responses
    • Rate limit headers in responses
    • Account is marked unavailable for selection

How to Check Rate Limit Status:

# View account status including rate limits
ccflare --list

# Check logs for rate limit messages
cat /tmp/ccflare-logs/app.log | grep "rate limited"

# View rate limit reset times in the dashboard
curl http://localhost:8080/api/accounts | jq '.[] | {name, rate_limit_status, rate_limit_reset}'

Recovery Strategies

When an account is rate-limited:

  1. ccflare automatically rotates to the next available account
  2. Rate-limited accounts are marked with a reset timestamp
  3. Accounts automatically become available again after the reset time

Manual recovery steps:

  1. Add more accounts to your pool:

    ccflare --add-account account2

  2. Check rate limit reset times in the dashboard:

    http://localhost:8080

  3. Monitor account-specific rate limits:

    # View rate limit details for each account
ccflare --list
# Look for rate_limit_status and rate_limit_reset columns

  4. Pause/resume accounts as needed:

    ccflare --pause <account-name>
ccflare --resume <account-name>

Connection Problems

Network Timeouts

Symptom: Requests hang or timeout

Error Messages:

  • ECONNREFUSED
  • ETIMEDOUT
  • Failed to forward unauthenticated request
  • All accounts failed to proxy the request

Solutions:

  1. Check your internet connection
  2. Verify the Anthropic API is accessible: 
    curl -I https://api.anthropic.com/v1/messages
  3. Check proxy settings if behind a corporate firewall
  4. Increase retry configuration in config file

Proxy Configuration

Configuring HTTP proxy:

# Set proxy environment variables before starting
export HTTP_PROXY=http://proxy.company.com:8080
export HTTPS_PROXY=http://proxy.company.com:8080
bun start

Bypass proxy for local requests:

export NO_PROXY=localhost,127.0.0.1

Performance Issues

Slow Responses

Symptoms:

  • High response times in logs
  • Dashboard shows increased latency

Solutions:

  1. Check account distribution:

    • Ensure accounts aren't all rate-limited
    • Verify load balancing strategy is appropriate

  2. Optimize retry settings:

    {
  "retry_attempts": 2,
  "retry_delay_ms": 500,
  "retry_backoff": 1.5
}

  3. Use session-based routing for conversational workloads:

    # Set strategy to session for better performance with conversations
# Session is the default and only supported strategy

High Memory Usage

Symptom: Process consuming excessive memory

Solutions:

  1. Check log file size (auto-rotates at 10MB):

    ls -lh /tmp/ccflare-logs/app.log

  2. Clear request history:

    ccflare --clear-history

  3. Restart the server to clear in-memory caches:

    # Graceful shutdown with Ctrl+C
# Then restart
bun start
# Or
ccflare --serve

Account Management Issues

Account Not Being Used

Symptom: Specific account never receives requests

Check:

  1. Account status:

    ccflare --list
# Look for: paused, rate_limited, or expired

  2. Session persistence:

    • Check if account has an active session
    • Verify session hasn't expired

Solutions:

  1. Resume the account if paused: 
    ccflare --resume <account-name>
  2. Wait for rate limit to reset
  3. Re-add the account if expired

Account Selection Problems

Symptom: Uneven distribution of requests

Solutions:

  1. Check current strategy:

    # Session strategy is the default and only supported strategy

  2. Session strategy behavior:

    • session: Maintains 1-hour sessions with individual accounts (default: 3600000ms)
    • This is the only supported strategy to avoid account bans
    • Adjust session_duration_ms if needed

Configuration Problems

Config File Location

Default locations by platform:

  • macOS: ~/.config/ccflare/ccflare.json
  • Linux: ~/.config/ccflare/ccflare.json
  • Windows: %LOCALAPPDATA%\ccflare\ccflare.json or %APPDATA%\ccflare\ccflare.json

Invalid Configuration

Symptom: Server fails to start or uses default values

Error Messages:

  • Failed to parse config file
  • Invalid strategy: [name]

Solutions:

  1. Validate JSON syntax:

    cat ~/.config/ccflare/ccflare.json | jq .

  2. Reset to defaults:

    # Backup current config
cp ~/.config/ccflare/ccflare.json ~/.config/ccflare/config.backup.json
# Remove corrupted config
rm ~/.config/ccflare/ccflare.json
# Restart server to create new config
bun start

Environment Variable Override

Environment variables override config file settings:

  • CLIENT_ID: OAuth client ID
  • PORT: Server port (default: 8080)
  • LB_STRATEGY: Load balancing strategy
  • RETRY_ATTEMPTS: Number of retry attempts
  • RETRY_DELAY_MS: Initial retry delay
  • SESSION_DURATION_MS: Session duration for session strategy

Database Issues

Database Path Problems

Symptom: Server fails to start with database errors

Error Messages:

  • Database file not found
  • Permission denied
  • Cannot create database

Solutions:

  1. Check database file permissions:

    # macOS/Linux
ls -la ~/.config/ccflare/ccflare.db

# Windows
dir %LOCALAPPDATA%\ccflare\ccflare.db

  2. Create the directory if it doesn't exist:

    # macOS/Linux
mkdir -p ~/.config/ccflare

# Windows
mkdir %LOCALAPPDATA%\ccflare

  3. Use a custom database path:

    export ccflare_DB_PATH=/path/to/custom/ccflare.db
bun start

Database Migration Failures

Symptom: Server logs show migration errors

Error Messages:

  • ALTER TABLE failed
  • Column already exists
  • Migration failed

Solutions:

  1. The migration system is idempotent - errors about existing columns are harmless

  2. If migrations fail repeatedly:

    # Backup existing database
cp ~/.config/ccflare/ccflare.db ~/.config/ccflare/ccflare.db.backup

# Remove and let it recreate
rm ~/.config/ccflare/ccflare.db
bun start

  3. Check for database corruption:

    sqlite3 ~/.config/ccflare/ccflare.db "PRAGMA integrity_check;"

Async Database Writer Issues

Symptom: Database writes appear delayed or missing

Error Messages:

  • Failed to execute DB job
  • Async DB writer queue flushed

Solutions:

  1. The async writer batches writes every 100ms for performance
  2. During shutdown, ensure graceful termination (Ctrl+C) to flush pending writes
  3. Check logs for async writer errors: 
    grep "async-db-writer" /tmp/ccflare-logs/app.log

Database Lock Errors

Symptom: Multiple processes accessing the database

Error Messages:

  • database is locked
  • SQLITE_BUSY

Solutions:

  1. Ensure only one instance of ccflare is running:

    ps aux | grep "bun start" | grep -v grep
ps aux | grep "ccflare --serve" | grep -v grep

  2. Kill any zombie processes:

    pkill -f "bun start"
pkill -f "ccflare --serve"

  3. Check for hanging database connections:

    lsof ~/.config/ccflare/ccflare.db

Streaming and Analytics Issues

Streaming Response Capture Problems

Symptom: Analytics data missing for streaming responses

Error Messages:

  • Stream tee error
  • Failed to capture streaming response
  • Buffer truncated at 1MB

Solutions:

  1. Streaming responses are captured up to 1MB for analytics
  2. Large responses will be truncated but still forwarded completely to the client
  3. Check if streaming is working: 
    # Look for streaming response logs
grep "Streaming response" /tmp/ccflare-logs/app.log

Analytics Data Issues

Symptom: Dashboard shows incorrect or missing analytics

Error Messages:

  • Failed to fetch analytics data
  • Analytics error:

Solutions:

  1. Check if requests are being recorded:

    # Count recent requests in database
sqlite3 ~/.config/ccflare/ccflare.db "SELECT COUNT(*) FROM requests WHERE timestamp > strftime('%s', 'now', '-1 hour') * 1000;"

  2. Verify analytics endpoint:

    # Test analytics API
curl "http://localhost:8080/api/analytics?range=1h"

  3. Clear and rebuild analytics data:

    ccflare --clear-history

  4. Reset account statistics without clearing history:

    ccflare --reset-stats

Usage Tracking Problems

Symptom: Token usage and costs not updating

Solutions:

  1. Usage is extracted from response headers and streaming data

  2. Check for usage extraction errors:

    grep "extractUsageInfo" /tmp/ccflare-logs/app.log

  3. Verify model pricing data:

    # Pricing updates every 24 hours by default
grep "Fetching latest pricing" /tmp/ccflare-logs/app.log

Logging and Debugging

Log File Locations

Logs are stored in the system's temporary directory:

  • All platforms: /tmp/ccflare-logs/app.log
  • Windows: %TEMP%\ccflare-logs\app.log

Enabling Debug Mode

Method 1: Environment Variable

export ccflare_DEBUG=1
export LOG_LEVEL=DEBUG
bun start

Method 2: Verbose Logging

# View real-time logs
tail -f /tmp/ccflare-logs/app.log

Log Formats

JSON Format (for parsing):

export LOG_FORMAT=json
bun start

Pretty Format (default):

[2024-01-20T10:30:45.123Z] INFO: [Proxy] Request completed for account1: 200 in 1234ms

Reading Logs

Filter by log level:

# View only errors
grep "ERROR" /tmp/ccflare-logs/app.log

# View warnings and errors
grep -E "WARN|ERROR" /tmp/ccflare-logs/app.log

Filter by component:

# View only proxy logs
grep "\[Proxy\]" /tmp/ccflare-logs/app.log

# View only server logs
grep "\[Server\]" /tmp/ccflare-logs/app.log

Common Error Messages

Authentication and Token Errors

"No active accounts available - forwarding request without authentication"

Meaning: All accounts are either paused, rate-limited, or expired

Solution:

  • Add new accounts or wait for rate limits to reset
  • Check account status: ccflare --list
  • Requests will be forwarded without authentication (may fail)

"Refresh promise not found for account"

Meaning: Internal error during token refresh process

Solution:

  • Restart the server to clear refresh state
  • Check for concurrent refresh attempts in logs

"Failed to refresh access token"

Meaning: OAuth refresh token is invalid or revoked

Solution:

  • Remove and re-add the account
  • Check if the OAuth app still has permissions in Anthropic console

Request Processing Errors

"Provider cannot handle path"

Meaning: Request path doesn't match expected Anthropic API patterns

Solution:

  • Ensure requests are to /v1/* endpoints
  • Check if you're using the correct base URL
  • Valid paths: /v1/messages, /v1/complete, etc.

"All accounts failed to proxy the request"

Meaning: Every account attempted but all failed

Response:

{
  "error": "All accounts failed to proxy the request",
  "attempts": 3,
  "lastError": "Error message here"
}

Solutions:

  1. Check individual account errors in logs
  2. Verify network connectivity
  3. Ensure at least one account has valid credentials
  4. Check for API outages

"Failed to forward unauthenticated request"

Meaning: Request forwarding without authentication failed

Solutions:

  1. Check network connectivity
  2. Verify Anthropic API is accessible
  3. Check for proxy configuration issues
  4. Look for timeout errors in logs

"Failed to proxy request with account"

Meaning: Error occurred while proxying with a specific account

Solutions:

  1. Check the specific account's status
  2. Look for token expiration or rate limits
  3. Verify network connectivity

Database Errors

"Failed to execute DB job"

Meaning: Async database write failed

Solutions:

  1. Check disk space
  2. Verify database file permissions
  3. Look for detailed error in logs

"database is locked"

Meaning: Another process is accessing the database

Solutions:

  1. Ensure only one ccflare instance is running
  2. Kill any zombie processes
  3. Wait for current operations to complete

Analytics and Streaming Errors

"Stream tee error"

Meaning: Failed to capture streaming response for analytics

Solutions:

  1. This doesn't affect the actual response to the client
  2. Check for memory issues if frequent
  3. Large responses may exceed 1MB capture limit

"Failed to fetch analytics data"

Meaning: Analytics query failed

Solutions:

  1. Check if database is accessible
  2. Verify time range parameters
  3. Clear history if data is corrupted: ccflare --clear-history

Configuration Errors

"Failed to parse config file"

Meaning: JSON syntax error in config file

Solutions:

  1. Validate JSON syntax: cat ~/.config/ccflare/ccflare.json | jq .
  2. Check for trailing commas or missing quotes
  3. Reset to defaults by deleting config file

"Invalid strategy: [name]"

Meaning: Unknown load balancing strategy specified

Solutions:

  1. Only valid strategy: session
  2. Check spelling in config or environment variable
  3. The default is already session

HTTP Status Codes

400 Bad Request

Common Causes:

  • Invalid request format
  • Missing required parameters
  • Invalid account name or ID

401 Unauthorized

Common Causes:

  • Expired access token
  • Invalid OAuth credentials
  • No active accounts available

403 Forbidden

Common Causes:

  • Account doesn't have required permissions
  • OAuth app restrictions

429 Too Many Requests

Common Causes:

  • Account rate limited
  • All accounts exhausted
  • Check rate limit headers for reset time

500 Internal Server Error

Common Causes:

  • Unexpected server error
  • Database connection issues
  • Check logs for stack trace

Startup Errors

"Address already in use"

Meaning: Port is already occupied

Solutions:

  1. Check if another instance is running: lsof -i :8080
  2. Use a different port: PORT=3000 bun start
  3. Kill the process using the port

"Cannot create database"

Meaning: Unable to create or access database file

Solutions:

  1. Check directory permissions
  2. Ensure parent directory exists
  3. Use custom path: export ccflare_DB_PATH=/custom/path/db.db

Environment Variables Reference

Core Configuration

VariableDescriptionDefaultExample
CLIENT_IDOAuth client ID for AnthropicNonemy-oauth-client-id
PORTServer port80803000
LB_STRATEGYLoad balancing strategysessionOnly session is supported
RETRY_ATTEMPTSNumber of retry attempts35
RETRY_DELAY_MSInitial retry delay in ms1000500
RETRY_BACKOFFRetry backoff multiplier21.5
SESSION_DURATION_MSSession duration for session strategy3600000 (1 hour)1800000

Paths and Storage

VariableDescriptionDefaultExample
ccflare_CONFIG_PATHCustom config file locationPlatform-specific/opt/ccflare/config.json
ccflare_DB_PATHCustom database locationPlatform-specific/opt/ccflare/data.db

Logging and Debugging

VariableDescriptionDefaultExample
ccflare_DEBUGEnable debug mode01
LOG_LEVELLog levelINFODEBUG, WARN, ERROR
LOG_FORMATLog formatprettyjson

Proxy Settings

VariableDescriptionDefaultExample
HTTP_PROXYHTTP proxy URLNonehttp://proxy.company.com:8080
HTTPS_PROXYHTTPS proxy URLNonehttp://proxy.company.com:8080
NO_PROXYBypass proxy for hostsNonelocalhost,127.0.0.1

Advanced Settings

VariableDescriptionDefaultExample
CF_PRICING_REFRESH_HOURSHours between pricing updates2412

Usage Examples

# Development setup with debug logging
export ccflare_DEBUG=1
export LOG_LEVEL=DEBUG
export LOG_FORMAT=json
bun start

# Production setup with custom paths
export ccflare_CONFIG_PATH=/etc/ccflare/config.json
export ccflare_DB_PATH=/var/lib/ccflare/data.db
export PORT=3000
bun start

# Corporate proxy setup
export HTTP_PROXY=http://proxy.corp.com:8080
export HTTPS_PROXY=http://proxy.corp.com:8080
export NO_PROXY=localhost,127.0.0.1,internal.corp.com
bun start

FAQ

Q: How do I know if ccflare is working?

A: Check the health endpoint:

curl http://localhost:8080/health

Expected response:

{
  "status": "ok",
  "accounts": {
    "total": 3,
    "active": 2,
    "paused": 1
  },
  "uptime": 3600000
}

Q: Can I use ccflare with multiple client applications?

A: Yes, ccflare acts as a transparent proxy. Point any Claude API client to http://localhost:8080 instead of https://api.anthropic.com.

Q: How do I backup my accounts?

A: The account data is stored in the SQLite database. Backup locations:

  • macOS/Linux: ~/.config/ccflare/ccflare.db
  • Windows: %LOCALAPPDATA%\ccflare\ccflare.db or %APPDATA%\ccflare\ccflare.db

Q: What happens during a graceful shutdown?

A: When receiving SIGINT (Ctrl+C) or SIGTERM:

  1. Stops accepting new requests
  2. Waits for in-flight requests to complete (with timeout)
  3. Flushes async database writer queue
  4. Closes database connections
  5. Flushes logs to disk
  6. Exits cleanly

Q: How do I migrate to a new machine?

A: Copy these files to the new machine:

  1. Database file (ccflare.db)
  2. Config file (ccflare.json)
  3. Set the same CLIENT_ID environment variable
  4. Ensure Bun is installed on the new machine

Q: Why is my analytics data missing or incorrect?

A: Several reasons can cause analytics issues:

  1. Streaming responses are only captured up to 1MB
  2. Database writes are async and may be delayed
  3. Usage data depends on response headers from Anthropic
  4. Check if requests are being recorded: sqlite3 ~/.config/ccflare/ccflare.db "SELECT COUNT(*) FROM requests;"

Q: How do I handle rate limits effectively?

A: Best practices for rate limit handling:

  1. Add multiple accounts to your pool
  2. Maintain proper session duration (default 1 hour)
  3. Monitor rate limit warnings in logs
  4. Set up alerts for rate-limited accounts
  5. Consider implementing request queuing in your application

Q: Can I use ccflare in production?

A: Yes, with these considerations:

  1. Use environment variables for sensitive configuration
  2. Set up proper logging and monitoring
  3. Use a persistent database path (not /tmp)
  4. Configure appropriate retry settings
  5. Add sufficient accounts for your load
  6. Use systemd or similar for process management

Q: Why are some accounts not being used?

A: Accounts may be skipped for several reasons:

  1. Paused: Manually paused via CLI
  2. Rate Limited: Temporarily unavailable due to rate limits
  3. Expired Token: Needs re-authentication
  4. Session: Account may have an active session
  5. Check status: ccflare --list

Q: How do I troubleshoot slow responses?

A: Steps to diagnose performance issues:

  1. Check response times in logs or analytics
  2. Verify no accounts are rate limited
  3. Look for retry attempts in logs
  4. Consider session duration settings
  5. Check network latency to Anthropic API
  6. Monitor database performance

Q: What's the difference between running with bun start vs ccflare --serve?

A: Both commands start the server:

  • bun start: Runs the server using the npm script
  • ccflare --serve: Runs the server directly via the CLI binary
  • Both are functionally equivalent

Getting Help

Reporting Bugs

When reporting issues, include:

  1. System Information:

    bun --version
node --version
echo $OSTYPE

  2. Error Logs:

    # Last 100 lines of logs
tail -n 100 /tmp/ccflare-logs/app.log

  3. Configuration (sanitized):

    # Remove sensitive data before sharing
cat ~/.config/ccflare/ccflare.json | jq 'del(.client_id)'

  4. Steps to Reproduce:

    • Exact commands run
    • Expected behavior
    • Actual behavior

Debug Information Script

Save this as debug-info.sh:

#!/bin/bash
echo "=== ccflare Debug Info ==="
echo "Date: $(date)"
echo "System: $(uname -a)"
echo "Bun Version: $(bun --version)"
echo "Node Version: $(node --version 2>/dev/null || echo 'Node not installed')"
echo ""

echo "=== Environment Variables ==="
env | grep -E "ccflare|CLIENT_ID|PORT|LB_STRATEGY|LOG_|PROXY" | sort
echo ""

echo "=== Process Info ==="
ps aux | grep -E "bun start|ccflare" | grep -v grep
echo ""

echo "=== Port Check ==="
lsof -i :${PORT:-8080} 2>/dev/null || echo "Port ${PORT:-8080} not in use"
echo ""

echo "=== Database Info ==="
if [ -f "$HOME/.config/ccflare/ccflare.db" ]; then
    echo "Database size: $(du -h "$HOME/.config/ccflare/ccflare.db" | cut -f1)"
    echo "Request count: $(sqlite3 "$HOME/.config/ccflare/ccflare.db" "SELECT COUNT(*) FROM requests;" 2>/dev/null || echo "Could not query")"
    echo "Account count: $(sqlite3 "$HOME/.config/ccflare/ccflare.db" "SELECT COUNT(*) FROM accounts;" 2>/dev/null || echo "Could not query")"
else
    echo "Database not found at default location"
fi
echo ""

echo "=== Recent Errors (last 24h) ==="
if [ -f "/tmp/ccflare-logs/app.log" ]; then
    grep "ERROR" /tmp/ccflare-logs/app.log | tail -20
else
    echo "Log file not found"
fi
echo ""

echo "=== Recent Rate Limits ==="
if [ -f "/tmp/ccflare-logs/app.log" ]; then
    grep -E "rate.?limit" /tmp/ccflare-logs/app.log | tail -10
else
    echo "Log file not found"
fi
echo ""

echo "=== Account Status ==="
ccflare --list 2>/dev/null || echo "Could not get account list"
echo ""

echo "=== API Health Check ==="
curl -s http://localhost:${PORT:-8080}/health | jq . 2>/dev/null || echo "Health check failed"
echo ""

echo "=== Recent Analytics (1h) ==="
curl -s "http://localhost:${PORT:-8080}/api/analytics?range=1h" | jq '.overview' 2>/dev/null || echo "Analytics unavailable"

Make the script executable:

chmod +x debug-info.sh
./debug-info.sh > debug-report.txt

Community Support

  • Check existing issues in the repository
  • Review this troubleshooting guide
  • Search logs for specific error messages
  • Try running in debug mode for more details

Performance Monitoring

Monitor key metrics via the dashboard API:

# Get current stats
curl http://localhost:8080/api/stats | jq .

# Get request history with filters
curl "http://localhost:8080/api/requests?limit=10&status=error" | jq .

# Get analytics with time ranges
curl "http://localhost:8080/api/analytics?range=1h" | jq .
curl "http://localhost:8080/api/analytics?range=24h" | jq .
curl "http://localhost:8080/api/analytics?range=7d" | jq .

# Get analytics with filters
curl "http://localhost:8080/api/analytics?range=1h&model=claude-3-opus&status=success" | jq .

# Monitor real-time logs
tail -f /tmp/ccflare-logs/app.log | grep -E "INFO|WARN|ERROR"

Quick Troubleshooting Checklist

When experiencing issues, check these in order:

  1. Service Health

    curl http://localhost:8080/health

  2. Account Status

    ccflare --list

  3. Recent Errors

    grep ERROR /tmp/ccflare-logs/app.log | tail -20

  4. Rate Limits

    grep "rate.?limit" /tmp/ccflare-logs/app.log | tail -10

  5. Network Connectivity

    curl -I https://api.anthropic.com/v1/messages

  6. Database Health

    sqlite3 ~/.config/ccflare/ccflare.db "PRAGMA integrity_check;"

Common Quick Fixes

ProblemQuick Fix
All accounts rate limitedAdd more accounts: ccflare --add-account newaccount
Token expiredRe-authenticate: ccflare --remove account && ccflare --add-account account
Database lockedKill duplicate processes: pkill -f "bun start"
Port in useUse different port: PORT=3000 bun start
Config corruptedReset config: rm ~/.config/ccflare/ccflare.json
Analytics missingClear history: ccflare --clear-history
Slow responsesCheck session duration settings (default 1 hour)

Remember: Most issues can be resolved by checking logs, verifying account status, and ensuring proper network connectivity. When in doubt, restart the service with debug logging enabled: ccflare_DEBUG=1 LOG_LEVEL=DEBUG bun start